One-Sample T-Test: Research Questions
One-Sample T-Test: Research Questions
One-Sample T-Test: Research Questions
com
Chapter 205
One-Sample T-Test
Introduction
This procedure provides several reports for making inference about a population mean based on a single sample.
These reports include confidence intervals of the mean or median, the t-test, the z-test, and non-parametric tests
including the randomization test, the quantile (sign) test, and the Wilcoxon Signed-Rank test. Tests of
assumptions and distribution plots are also available in this procedure.
Another procedure that produces a large amount of summary information about a single sample is the Descriptive
Statistics procedure. While it is not as focused on hypothesis testing, it contains many additional descriptive
statistics, including minimum, maximum, range, counts, trimmed means, sums, mode, variance, Skewness,
Kurtosis, coefficient of variation, coefficient of dispersion, percentiles, additional normality tests, and a stem-and-
leaf plot.
Research Questions
For the one-sample situation, the typical concern in research is examining a measure of central tendency
(location) for the population of interest. The best-known measures of location are the mean and median. For a
one-sample situation, we might want to know if the average waiting time in a doctor’s office is greater than one
hour, if the average refund on a 1040 tax return is different from $500, if the average assessment for similar
residential properties is less than $120,000, or if the average growth of roses is 4 inches or more after two weeks
of treatment with a certain fertilizer.
One early concern should be whether the data are normally distributed. If normality can safely be assumed, then
the one-sample t-test is the best choice for assessing whether the measure of central tendency, the mean, is
different from a hypothesized value. On the other hand, if normality is not valid, one of the nonparametric tests,
such as the Wilcoxon Signed Rank test or the quantile test, can be applied.
Technical Details
The technical details and formulas for the methods of this procedure are presented in line with the Example 1
output. The output and technical details are presented in the following order:
• Descriptive Statistics
• Confidence Interval of 𝜇𝜇 with 𝜎𝜎 Unknown
• Confidence Interval of 𝜇𝜇 with 𝜎𝜎 Known
• Confidence Interval of the Median
• Bootstrap Confidence Intervals
• Confidence Interval of 𝜎𝜎
• T-Test and associated power report
• Z-Test
205-1
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
One-Sample T-Test
• Randomization Test
• Quantile (Sign) Test
• Wilcoxon Signed-Rank Test
• Tests of Assumptions
• Graphs
Data Structure
For this procedure, the data are entered as a single column and specified as a response variable. Multiple columns
can be analyzed individually during the same run if multiple response variables are specified.
Weight
159
155
157
125
103
122
101
82
228
199
195
110
191
151
119
119
112
87
190
87
205-2
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
One-Sample T-Test
Assumptions
This section describes the assumptions that are made when you use one of these tests. The key assumption relates
to normality or nonnormality of the data. One of the reasons for the popularity of the t-test is its robustness in the
face of assumption violation. However, if an assumption is not met even approximately, the significance levels
and the power of the t-test are invalidated. Unfortunately, in practice it often happens that more than one
assumption is not met. Hence, take the steps to check the assumptions before you make important decisions based
on these tests. There are reports in this procedure that permit you to examine the assumptions, both visually and
through assumptions tests.
Procedure Options
This section describes the options available in this procedure.
Variables Tab
This option specifies the variables that will be used in the analysis.
Response Variables
Response Variable(s)
Specify one or more columns for one-sample analysis. A separate one-sample report will be output for each
column listed.
205-4
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
One-Sample T-Test
Reports Tab
The options on this panel specify which reports will be included in the output.
Descriptive Statistics
Descriptive Statistics
This section reports the count, mean, standard deviation, standard error, and mean for the specified variable.
Confidence Intervals
Confidence Level
This confidence level is used for the unknown σ and known σ confidence intervals of the mean, as well as for the
confidence intervals of the median and standard deviation. Typical confidence levels are 90%, 95%, and 99%,
with 95% being the most common.
Limits
Specify whether a two-sided or one-sided confidence interval of the mean or standard deviation is to be reported.
Confidence intervals of the median and bootstrap confidence intervals are always two-sided in these reports.
• Two-Sided
For this selection, the lower and upper limits of 𝜇𝜇are reported, giving a confidence interval of the form (Lower
Limit, Upper Limit).
• One-Sided Upper
For this selection, only an upper limit of 𝜇𝜇 is reported, giving a confidence interval of the form (-∞, Upper
Limit).
• One-Sided Lower
For this selection, only a lower limit of 𝜇𝜇 is reported, giving a confidence interval of the form (Lower Limit,
∞).
205-5
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
One-Sample T-Test
Examples:
90 95 99
90:99(1)
90
Samples (N)
This is the number of bootstrap samples used. A general rule of thumb is to use at least 100 when standard errors
are the focus or 1000 when confidence intervals are your focus. With current computer speeds, 10,000 or more
samples can usually be run in a short time.
Retries
If the results from a bootstrap sample cannot be calculated, the sample is discarded and a new sample is drawn in
its place. This parameter is the number of times that a new sample is drawn before the algorithm is terminated. It
is possible that the sample cannot be bootstrapped. This parameter lets you specify how many alternative
bootstrap samples are considered before the algorithm gives up. The range is 1 to 1000, with 50 as a
recommended value.
Bootstrap Confidence Interval Method
This option specifies the method used to calculate the bootstrap confidence intervals.
• Percentile
The confidence limits are the corresponding percentiles of the bootstrap values.
• Reflection
The confidence limits are formed by reflecting the percentile limits. If X0 is the original value of the
parameter estimate and XL and XU are the percentile confidence limits, the Reflection interval is (2 X0 - XU,
2 X0 - XL).
Percentile Type
This option specifies which of five different methods is used to calculate the percentiles. More details of these
five types are offered in the Descriptive Statistics chapter.
Confidence Interval of 𝝈𝝈
This section provides one- or two-sided confidence limits for σ. These limits rely heavily on the assumption of
normality of the population from which the sample is taken.
Tests
Alpha
Alpha is the significance level used in the hypothesis tests. A value of 0.05 is most commonly used, but 0.1,
0.025, 0.01, and other values are sometimes used. Typical values range from 0.001 to 0.20.
H0 Value
This is the hypothesized value of the population mean under the null hypothesis. This is the mean to which the
sample mean will be compared.
205-6
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
One-Sample T-Test
Tests - Parametric
T-Test
This report provides the results of the common one-sample T-Test.
Power Report for T-Test
This report gives the power of the one-sample T-Test when it is assumed that the population mean and standard
deviation are equal to the sample mean and standard deviation.
Z-Test
This report gives the Z-Test for comparing a sample mean to a hypothesized mean assuming a known standard
deviation. This procedure is rarely used in practice because the population standard deviation is rarely known.
Known Standard Deviation 𝛔𝛔
This is the known population standard deviation used in the corresponding Z-Test.
Tests - Nonparametric
Randomization Test
A randomization test is conducted by first determining the signs of all the values relative to the null hypothesized
mean – that is, the signs of the values after subtracting the null hypothesized mean. Then all possible
permutations of the signs are enumerated. Each of these permutations is assigned to absolute values of the original
subtracted values in their original order, and the corresponding t-statistic is calculated. The original t-statistic,
based on the original data, is compared to each of the permutation t-statistics, and the total count of permutation t-
statistics more extreme than the original t-statistic is determined. Dividing this count by the number of
permutations tried gives the significance level of the test.
For even moderate sample sizes, the total number of permutations is in the trillions, so a Monte Carlo approach is
used in which the permutations are found by random selection rather than complete enumeration. Edgington
suggests that at least 1,000 permutations be selected. We suggest that this be increased to 10,000.
Monte Carlo Samples
Specify the number of Monte Carlo samples used when conducting randomization tests. You also need to check
the ‘Randomization Test’ box under the Variables tab to run this test.
Somewhere between 1000 and 100000 Monte Carlo samples are usually necessary. Although the default is 1000,
we suggest the use of 10000 when using this test.
Quantile (Sign) Test
The quantile (sign) test tests whether the null hypothesized quantile value is the quantile for the quantile test
proportion.
Each of the values in the sample is compared to the null hypothesized quantile value and determined whether it is
larger or smaller. Values equal to the null hypothesized value are removed. The binomial distribution based on the
Quantile Test Proportion is used to determine the probabilities in each direction of obtaining such a sample or one
more extreme if the null quantile is the true one. These probabilities are the p-values for the one-sided test of each
direction. The two-sided p-value is twice the value of the smaller of the two one-sided p-values.
205-7
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
One-Sample T-Test
When the quantile of interest is the median (the quantile test proportion is 0.5), this quantile test is called the sign
test.
Quantile Test Proportion
This is the value of the binomial proportion used in the Quantile test. The quantile (sign) test tests whether the
null hypothesized quantile value is the quantile for this proportion. A value of 0.5 results in the Sign Test.
Wilcoxon Signed-Rank Test
This nonparametric test makes use of the sign and the magnitude of the rank of the differences (sample values
minus the hypothesized value). It is one of the most commonly used nonparametric alternatives to the one-sample
t-test.
There are 3 different tests that can be conducted:
• Exact Test
The exact test can be calculated if there are no ties. This test is recommended when there are no ties.
Assumptions
Tests of Assumptions
This report gives a series of four tests of the normality assumptions.
• Shapiro-Wilk Normality
• Skewness Normality
• Kurtosis Normality
• Omnibus Normality
Assumptions Alpha
Assumptions Alpha is the significance level used in all the assumptions tests. A value of 0.05 is typically used for
hypothesis tests in general, but values other than 0.05 are often used for the case of testing assumptions. Typical
values range from 0.001 to 0.20.
205-8
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
One-Sample T-Test
Report Options
Variable Names
This option lets you select whether to display only variable names, variable labels, or both.
Decimal Places
Decimals
This option allows the user to specify the number of decimal places directly or based on the significant digits.
These decimals are used for output except the nonparametric tests and bootstrap results, where the decimal places
are defined internally.
If one of the significant digits options is used, the ending zero digits are not shown. For example, if 'Significant
Digits (Up to 7)' is chosen,
0.0500 is displayed as 0.05
1.314583689 is displayed as 1.314584
The output formatting system is not designed to accommodate 'Significant Digits (Up to 13)', and if chosen, this
will likely lead to lines that run on to a second line. This option is included, however, for the rare case when a
very large number of decimals is needed.
Plots Tab
The options on this panel control the inclusion and appearance of the plots.
Select Plots
Histogram … Average-Difference Plot
Check the boxes to display the plot. Click the plot format button to change the plot settings.
Bootstrap Confidence Interval Histograms
This plot requires that the bootstrap confidence intervals report has been selected. It gives the plot of the bootstrap
distribution used in creating the bootstrap confidence interval.
205-9
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
One-Sample T-Test
205-10
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
One-Sample T-Test
The following reports and charts will be displayed in the Output window.
Variable
The name of the variable for which the descriptive statistics are listed here.
Count
This is the number of non-missing values.
Mean
This is the average of the data values.
∑𝑛𝑛𝑖𝑖=1 𝑥𝑥𝑖𝑖
𝑥𝑥 =
𝑛𝑛
Standard Deviation
The sample standard deviation is the square root of the variance. It is a measure of dispersion based on squared
distances from the mean for the variables listed.
Standard Error
This is the estimated standard deviation of the distribution of sample means for an infinite population.
𝑠𝑠
𝑆𝑆𝑆𝑆𝑥𝑥 =
√𝑛𝑛
Median
The median is the 50th percentile of the data, using the AveXp(n+1) method. The details of this method are
described in the Descriptive Statistics chapter under Percentile Type.
Variable
This is the spreadsheet column variable.
Count
This is the number of non-missing values.
Mean
This is the average of the data values.
205-11
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
One-Sample T-Test
Standard Deviation
The sample deviation is the square root of the variance. It is a measure of dispersion based on squared distances
from the mean for the variables listed.
Standard Error
This is the estimated standard deviation of the distribution of sample means for an infinite population.
T*
This is the t-value used to construct the confidence limits. It is based on the degrees of freedom and the
confidence level.
DF
The degrees of freedom are used to determine the T distribution from which T* is generated:
𝑑𝑑𝑑𝑑 = 𝑛𝑛 − 1
Lower and Upper Confidence Limits
These are the confidence limits of the confidence interval for 𝜇𝜇. The confidence interval formula is
∗
𝑥𝑥 ± 𝑇𝑇𝑑𝑑𝑑𝑑 ∙ 𝑆𝑆𝑆𝑆𝑥𝑥
Variable
This is the spreadsheet column variable.
Count
This is the number of non-missing values.
Mean
This is the average of the data values.
205-12
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
One-Sample T-Test
Variable
This is the spreadsheet column variable.
Count
This is the number of non-missing values.
Median
The median is the 50th percentile of the data, using the AveXp(n+1) method. The details of this method are
described in the Descriptive Statistics chapter under Percentile Type.
Lower Limit and Upper Limit
These are the lower and upper confidence limits of the median. These limits are exact and make no distributional
assumptions other than a continuous distribution. No limits are reported if the algorithm for this interval is not
able to find a solution. This may occur if the number of unique values is small.
Bootstrap Section
Bootstrapping was developed to provide standard errors and confidence intervals in situations in which the
standard assumptions are not valid. In these nonstandard situations, bootstrapping is a viable alternative to the
corrective action suggested earlier. The method is simple in concept, but it requires extensive computation time.
The bootstrap is simple to describe. You assume that your sample is actually the population and you draw B
samples (B is over 1000) of N from the original dataset, with replacement. With replacement means that each
observation may be selected more than once. For each bootstrap sample, the mean is computed and stored.
Suppose that you want the standard error and a confidence interval of the mean. The bootstrap sampling process
has provided B estimates of the mean. The standard deviation of these B means is the bootstrap estimate of the
standard error of the mean. The bootstrap confidence interval is found by arranging the B values in sorted order
and selecting the appropriate percentiles from the list. For example, a 90% bootstrap confidence interval for the
difference is given by fifth and ninety-fifth percentiles of the bootstrap mean values.
The main assumption made when using the bootstrap method is that your sample approximates the population
fairly well. Because of this assumption, bootstrapping does not work well for small samples in which there is little
likelihood that the sample is representative of the population. Bootstrapping should only be used in medium to
large samples.
205-13
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
One-Sample T-Test
This report provides bootstrap confidence intervals of the mean. Note that since these results are based on 3000
random bootstrap samples, they will differ slightly from the results you obtain when you run this report.
Original Value
This is the parameter estimate obtained from the complete sample without bootstrapping.
Bootstrap Mean
This is the average of the parameter estimates of the bootstrap samples.
Bias (BM - OV)
This is an estimate of the bias in the original estimate. It is computed by subtracting the original value from the
bootstrap mean.
Bias Corrected
This is an estimated of the parameter that has been corrected for its bias. The correction is made by subtracting the
estimated bias from the original parameter estimate.
Standard Error
This is the bootstrap method’s estimate of the standard error of the parameter estimate. It is simply the standard
deviation of the parameter estimate computed from the bootstrap estimates.
Conf. Level
This is the confidence coefficient of the bootstrap confidence interval given to the right.
Bootstrap Confidence Limits – Lower and Upper
These are the limits of the bootstrap confidence interval with the confidence coefficient given to the left. These
limits are computed using the confidence interval method (percentile or reflection) designated on the Bootstrap
panel.
Note that to be accurate, these intervals must be based on over a thousand bootstrap samples and the original
sample must be representative of the population.
205-14
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
One-Sample T-Test
Bootstrap Histogram
The histogram shows the distribution of the bootstrap parameter estimates.
Variable
This is the spreadsheet column variable.
Count
This is the number of non-missing values.
Standard Deviation
The sample deviation is the square root of the variance. It is a measure of dispersion based on squared distances
from the mean for the variables listed.
T-Test Section
This section presents the results of the traditional one-sample T-test. Here, reports for all three alternative
hypotheses are shown, but a researcher would typically choose one of the three before generating the output. All
three tests are shown here for the purpose of exhibiting all the output options available.
One-Sample T-Test
Alternative Hypothesis
The (unreported) null hypothesis is
𝐻𝐻0 : 𝜇𝜇 = 𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻ℎ𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉
205-15
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
One-Sample T-Test
Reject H0 at α = (0.050)
This column indicates whether or not the null hypothesis is rejected, in favor of the alternative hypothesis, based
on the p-value and chosen α. A test in which the null hypothesis is rejected is sometimes called significant.
205-16
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
One-Sample T-Test
Alternative Hypothesis
This value identifies the test direction of the test reported in this row. In practice, you would select the alternative
hypothesis prior to your analysis and have only one row showing here.
N
N is the assumed sample size.
µ
This is the assumed population mean on which the power calculation is based.
σ
This is the assumed population standard deviation on which the power calculation is based.
Z-Test Section
This section presents the results of the traditional one-sample Z-test, which can be used when the population
standard deviation is known. Because the population standard deviation is rarely known, this test is not commonly
used in practice. Here, reports for all three alternative hypotheses are shown, but a researcher would typically
choose one of the three before generating the output. All three tests are shown here for the purpose of exhibiting
all the output options available.
One-Sample Z-Test
Alternative Hypothesis
The (unreported) null hypothesis is
𝐻𝐻0 : 𝜇𝜇 = 𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻ℎ𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉
205-17
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
One-Sample T-Test
Reject H0 at α = (0.050)
This column indicates whether or not the null hypothesis is rejected, in favor of the alternative hypothesis, based
on the p-value and chosen α. A test in which the null hypothesis is rejected is sometimes called significant.
205-18
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
One-Sample T-Test
Prob Reject H0
Test Level at α = 0.050?
Randomization Test 0.33700 No
Prob Level
The probability level, also known as the p-value or significance level, is the probability that the test statistic will
take a value at least as extreme as the observed value, assuming that the null hypothesis is true. If the p-value is
less than the prescribed α, in this case 0.05, the null hypothesis is rejected in favor of the alternative hypothesis.
Otherwise, there is not sufficient evidence to reject the null hypothesis.
Reject H0 at α = (0.050)
This column indicates whether or not the null hypothesis is rejected, in favor of the alternative hypothesis, based
on the p-value and chosen α. A test in which the null hypothesis is rejected is sometimes called significant.
based upon the sampling distribution of the statistic being normal under the alternative hypothesis.
205-19
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
One-Sample T-Test
205-20
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
One-Sample T-Test
Std Dev of W
This is the standard deviation of the sampling distribution of the sum of ranks. Here ti represents the number of
times the ith value occurs.
205-21
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
One-Sample T-Test
The main assumption when using the t-test is that the data are normally distributed. The normality assumption can
be checked statistically by the Shapiro-Wilk, skewness, kurtosis, or omnibus normality tests and visually by the
histogram or normal probability plot.
In the case of non-normality, one of the nonparametric tests may be considered. Sometimes a transformation, such
as the natural logarithm or the square root of the original data, can change the underlying distribution from
skewed to normal. To evaluate whether the underlying distribution of the variable is normal after the
transformation, rerun the normal probability plot on the transformed variable. If some of the data values are
negative or zero, it may be necessary to add a constant to the original data prior to the transformation. If the
transformation produces a suitable result, then the one-sample t-test could be performed on the transformed data.
Normality (Shapiro-Wilk, Skewness, Kurtosis, and Omnibus)
These four tests allow you to test the skewness, kurtosis, and overall normality of the data. If any of them reject
the hypothesis of normality, the data should not be considered normal. These tests are discussed in more detail in
the Descriptive Statistics chapter.
205-22
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
One-Sample T-Test
Histogram
The histogram provides a general idea of the shape. The histogram appearance can be affected considerably by the
bin width choices, particularly when the sample size is smaller.
Probability Plot
If the observations fall along a straight line, this indicates the data follow a normal distribution.
205-23
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
One-Sample T-Test
This average-difference plot is designed to detect a lack of symmetry in the data. This plot is constructed from the
original data. Let D(i) represent the ith ordered value. Pairs of these sorted values are considered, with the pairing
being done as you move toward the middle from either end. That is, consider the pairs D(1) and D(n), D(2) and
D(n-1), D(3) and D(n-2), etc. Plot the average versus the difference of each of these pairs. Your plot will have
about n/2 points, depending on whether n is odd or even. If the data are symmetric, the average of each pair will
be the close to the median.
Symmetry is an important assumption for some of the available tests. A perfectly symmetric set of data should
show a vertical line of points hitting the horizontal axis at the value of the median. Departures from symmetry
would deviate from this standard.
Introduction
This step involves scanning your data for anomalies, data entry errors, typos, and so on. Frequently we hear of
people who completed an analysis with the right techniques but obtained strange conclusions because they had
mistakenly selected the data.
Sample Size
The sample size (number of non-missing rows) has a lot of ramifications. The larger the sample size for the one-
sample t-test the better. Of course, the t-test may be performed on very small samples, say 4 or 5 observations, but
it is impossible to assess the validity of assumptions with such small samples. It is our statistical experience that at
least 20 observations are necessary to evaluate normality properly. On the other hand, since skewness can have
unpleasant effects on t-tests with small samples, particularly for one-tailed tests, larger sample sizes (30 to 50)
may be necessary.
205-24
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
One-Sample T-Test
It is possible to have a sample size that is too large for a statistical significance test. When your sample size is
very large, you are almost guaranteed to find statistical significance. However, the question that then arises is
whether the magnitude of the difference is of practical importance.
Missing Values
The number and pattern of missing values is always an issue to consider. Usually, we assume that missing values
occur at random throughout your data. If this is not true, your results will be biased since a particular segment of
the population is underrepresented. If you have a lot of missing values, some researchers recommend comparing
other variables with respect to missing versus non-missing. If you find large differences in other variables, you
should begin to worry about whether the missing values might cause a systematic bias in your results.
Type of Data
The mathematical basis of the t-test assumes that the data are continuous. Because of the rounding that occurs
when data are recorded, all data are technically discrete. The validity of assuming the continuity of the data then
comes down to determining when we have too much rounding. For example, most statisticians would not worry
about human-age data that was rounded to the nearest year. However, if these data were rounded to the nearest ten
years or further to only three groups (young, adolescent, and adult), most statisticians would question the validity
of the probability statements. Some studies have shown that the t-test is reasonably accurate when the data has
only five possible values (most would call this discrete data). If your data contains less than five unique values,
any probability statements made are tenuous.
Outliers
Generally, outliers cause distortion in statistical tests. You must scan your data for outliers (the box plot is an
excellent tool for doing this). If you have outliers, you have to decide if they are one-time occurrences or if they
would occur in another sample. If they are one-time occurrences, you can remove them and proceed. If you know
they represent a certain segment of the population, you have to decide between biasing your results (by removing
them) or using a nonparametric test that can deal with them. Most would choose the nonparametric test.
Introduction
NCSS is designed to be simple to operate, but it requires some learning. When you go to run a procedure such as
this for the first time, take a few minutes to read through the chapter again and familiarize yourself with the issues
involved.
Enter Variables
The NCSS panels are set with ready-to-run defaults. About all you have to do is select the appropriate variable.
Specify Alpha
Most beginners in statistics forget this important step and let the alpha value default to the standard 0.05. You
should make a conscious decision as to what value of alpha is appropriate for your study. The 0.05 default came
about when people had to rely on printed probability tables in which there were only two values available: 0.05 or
0.01. Now you can set the value to whatever is appropriate.
205-25
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
One-Sample T-Test
Introduction
Once the program output is displayed, you will be tempted to go directly to the probability of the t-test, determine
if you have a significant result, and proceed to something else. However, it is very important that you proceed
through the output in an orderly fashion. The first task is to determine which of the assumptions are met by your
data.
Sometimes, when the data are nonnormal, a data transformation (like square roots or logs) might normalize the
data. Frequently, this kind of transformation or re-expression approach works very well. However, always check
the transformed variable to see if it is normally distributed.
It is not unusual in practice to find a variety of tests being run on the same basic null hypothesis. That is, the
researcher who fails to reject the null hypothesis with the first test will sometimes try several others and stop
when the hoped-for significance is obtained. For instance, a statistician might run the one-sample t-test on the
original data, the one-sample t-test on the logarithmically transformed data, the Wilcoxon rank-sum test, and the
Quantile test. An article by Gans (1984) suggests that there is no harm on the true significance level if no more
than two tests are run. This is not a bad option in the case of questionable outliers. However, as a rule of thumb, it
seems more honest to investigate whether the data is normal. The conclusion from that investigation should direct
you to the right test.
Random Sample
The validity of this assumption depends on the method used to select the sample. If the method used ensures that
each individual in the population of interest has an equal probability of being selected for this sample, you have a
random sample. Unfortunately, you cannot tell if a sample is random by looking at either it or statistics from it.
Normality
To validate this assumption, you would first look at the plots. Outliers will show up on the box plots and the
probability plots. Skewness, kurtosis, more than one mode, and a host of other problems will be obvious from the
density trace on the histogram. After considering the plots, look at the Tests of Assumptions Section to get
numerical confirmation of what you see in the plots. Remember that the power of these normality tests is directly
related to the sample size, so when the normality assumption is accepted, double-check that your sample is large
enough to give conclusive results (at least 20).
Symmetry
The nonparametric tests need the assumption of symmetry. The easiest ways to evaluate this assumption are from
the density trace on the histogram or from the average-difference plot.
Introduction
After understanding how your data fit the assumptions of the various one-sample tests, you are ready to determine
which statistical procedures will be valid. You should select one of the following three situations based on the
status of the normality.
205-26
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
One-Sample T-Test
Normal Data
Use the T-Test Section for hypothesis testing and the Descriptive Statistics Section for interval estimation.
Introduction
You are now ready to conduct your test. Depending on the nature of your study, you should look at either of the
following sections.
Hypothesis Testing
Here you decide whether to use a two-tailed or one-tailed test. The two-tailed test is the standard. If the
probability level is less than your chosen alpha level, reject the null hypothesis of equality to a specified mean (or
median) and conclude that the mean is different. Your next task is to look at the mean itself to determine if the
size of the difference is of practical interest.
Confidence Limits
The confidence limits let you put bounds on the size of the mean (for one independent sample) or mean difference
(for dependent samples). If these limits are narrow and close to your hypothesized value, you might determine
that even though your results are statistically significant, there is no practical significance.
205-27
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
One-Sample T-Test
205-28
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
One-Sample T-Test
The normal probability plot above does not look straight. It shows some skewness to the right. The histogram
confirms the skewness to the right. This type of skewness to the right turns up quite often when dealing with
elapsed-time data.
The skewness, kurtosis, and the omnibus normality tests in the output below have p-values greater than 0.05,
indicating that answer time seems to be normally distributed. This conflict in conclusions between the normal
probability plot and the normality tests is probably due to the fact that this sample size is not large enough to
accurately assess the normality of the data.
One-Sample T-Test
205-29
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
One-Sample T-Test
205-30
© NCSS, LLC. All Rights Reserved.