# Paired and Independent t-Test in Excel for Action Research

How to run paired and independent t-tests in Excel for DepEd action research, step by step, with the toolkit's practice data, a CRLA example and effect sizes.

This is the Markdown copy of [Paired and Independent t-Test in Excel for Action Research](https://nash.com.ph/blog/t-test-excel-action-research), an article on the NASH Reviewer Blog by Mark Anthony Llego, published Oct 28, 2026, updated Nov 4, 2026. Topics: Action research, Learning outcomes. NASH Reviewer is an independent practice platform not affiliated with DepEd, BHROD, or NEAP.

The Department of Education's (DepEd) Action Research Toolkit teaches two statistical tests, the paired t-test and the independent t-test. Both run in Microsoft Excel, as two worked examples show: the toolkit's own practice exercise, the Math scores of 12 boys and 12 girls taught with games, and a real example from public data, the share of Grade 1 learners struggling in Mother Tongue reading in the 22 public schools of one division at the start and end of SY 2025-26.

It is written for teachers analyzing the pretest and posttest scores of their action research, for school heads and research coordinators who review the results, and for anyone checking their answers to the toolkit's exercise. The toolkit's answer key leaves the t-values and p-values blank and gives one critical value that does not match its degrees of freedom; the completed key and the correct critical values are below.

## Key points

- **Two designs, two tests.** Use the **paired t-test** when the same learners are measured twice (a pretest and a posttest). Use the **independent t-test** when two different groups are compared (an experimental and a control group).
- **Excel does both.** Turn on the Analysis ToolPak once, then use Data, Data Analysis. The `T.TEST` function gives the p-value directly: type 1 is paired, type 2 is two groups with equal variances, type 3 is two groups with unequal variances.
- **Read the two-tailed p-value.** Unless you stated a directional hypothesis before collecting data, use "P(T<=t) two-tail". Reject the null hypothesis when it is below 0.05.
- **The toolkit's exercise, solved.** Boys' scores rose (t(11) = 17.39, p < 0.001); girls' scores fell (t(11) = −3.25, p = 0.008); boys scored higher than girls on the posttest (t(22) = 8.85, p < 0.001). The correct two-tailed critical values are 2.201 for 11 degrees of freedom and 2.074 for 22, not 2.10.
- **Report the size, not only the p-value.** Add an effect size (Cohen's d) with a simple Excel formula, and the means and standard deviations.
- **A significant gain is not proof that the intervention worked.** In Candon City's schools, the share of Grade 1 struggling readers fell from a mean of 79.0% to 4.6% in one school year with no special intervention (t(21) = −16.26, p < 0.001). Ordinary growth passes the test too.

## Which t-test to use

| Your design | Example | Test | Excel Data Analysis tool | `T.TEST` type |
|:--|:--|:--|:--|--:|
| One group, measured before and after | Your section's pretest and posttest | Paired t-test | t-Test: Paired Two Sample for Means | 1 |
| Two different groups, similar spread of scores | Experimental and control sections' posttests | Independent t-test, equal variances | t-Test: Two-Sample Assuming Equal Variances | 2 |
| Two different groups, very different spread | Same, when one group's scores vary much more | Independent t-test, unequal variances (Welch) | t-Test: Two-Sample Assuming Unequal Variances | 3 |

The toolkit puts it this way: the independent t-test compares "means of two groups on the same variable" whose members "are not the same", and the paired t-test compares "two means obtained from the same group". A design with an experimental and a control group, each given a pretest and a posttest, uses both: a paired test within each group, and an independent test between the groups' posttests or, better, their gains.

## Set up the data

Put one learner per row. Use codes, not names; the toolkit's ethics protocols require that learners' identities stay confidential.

- **For a paired test:** a column for the learner's code, one for the pretest and one for the posttest. Each row must have both scores; leave out learners missing either one, and say so in your report.
- **For an independent test:** each group's scores in its own column. The groups may have different numbers of learners.

The toolkit's practice data, laid out this way (boys in rows 2 to 13, girls in rows 14 to 25):

| Code | Pretest | Posttest | Code | Pretest | Posttest |
|:--|--:|--:|:--|--:|--:|
| B1 | 8 | 15 | G1 | 9 | 6 |
| B2 | 5 | 12 | G2 | 7 | 7 |
| B3 | 7 | 16 | G3 | 8 | 4 |
| B4 | 4 | 10 | G4 | 6 | 5 |
| B5 | 10 | 18 | G5 | 11 | 6 |
| B6 | 7 | 15 | G6 | 8 | 8 |
| B7 | 4 | 9 | G7 | 9 | 4 |
| B8 | 7 | 14 | G8 | 10 | 8 |
| B9 | 10 | 15 | G9 | 8 | 5 |
| B10 | 7 | 13 | G10 | 7 | 6 |
| B11 | 9 | 15 | G11 | 5 | 7 |
| B12 | 6 | 11 | G12 | 8 | 6 |

![Two slope charts of the Action Research Toolkit's practice data, Math scores of 12 boys and 12 girls taught with games. Every boy's score rises, and the boys' mean goes from 7.00 to 13.58; most girls' scores fall, and the girls' mean goes from 8.00 to 6.00.](https://nash.com.ph/blog/_astro/toolkit-t-test-pretest-posttest.C6eEEUNV_Z7zhs5.webp)

*Figure 1. The Action Research Toolkit's practice data: each learner's Math pretest and posttest, with the group means. Chart: © 2026 Mark Anthony Llego, NASH Reviewer. [Download the full-resolution chart (PNG)](https://nash.com.ph/blog/charts/deped-data/toolkit-t-test-pretest-posttest.png).*

Always look at the data before testing it. The chart already tells the story: every boy improved, and most girls' scores went down.

## Turn on the Analysis ToolPak

The Analysis ToolPak is a free add-in that comes with Excel. Microsoft's steps:

1. On Windows, select **File**, **Options**, then **Add-ins**. In the **Manage** box at the bottom, choose **Excel Add-ins** and select **Go**. On a Mac, open **Tools** and select **Excel Add-ins**.
2. Check **Analysis ToolPak** and select **OK**.
3. A **Data Analysis** button now appears on the **Data** tab. If it does not, restart Excel.

You only do this once. The `T.TEST` function works without the add-in, and also in Google Sheets.

## The paired t-test, step by step

The toolkit's first two questions ask whether the boys' and the girls' Math scores changed from pretest to posttest.

**1. State the hypotheses.** Null hypothesis: there is no difference between the boys' pretest and posttest Math scores. Research hypothesis: there is a difference. These are nondirectional, so the test is two-tailed.

**2. Choose the level of significance.** Alpha = 0.05, as the toolkit uses.

**3. Run the test.** Select **Data**, **Data Analysis**, **t-Test: Paired Two Sample for Means**, then **OK**. Fill in the dialog:

- **Variable 1 Range:** the posttest scores (C2:C13). Putting the posttest first makes the t-value positive when scores rise.
- **Variable 2 Range:** the pretest scores (B2:B13).
- **Hypothesized Mean Difference:** 0.
- **Labels:** check it only if you selected the header cells too.
- **Alpha:** 0.05.
- **Output options:** a cell on the sheet, or a new worksheet.

**4. Read the output.** For the boys, Excel prints:

|                              | Posttest     | Pretest     |
|:-----------------------------|:-------------|:------------|
| Mean                         | 13.58333333  | 7           |
| Variance                     | 6.992424242  | 4.181818182 |
| Observations                 | 12           | 12          |
| Pearson Correlation          | 0.8742068402 |             |
| Hypothesized Mean Difference | 0            |             |
| df                           | 11           |             |
| t Stat                       | 17.39043739  |             |
| P(T<=t) one-tail             | 1.19073E-09  |             |
| t Critical one-tail          | 1.795884819  |             |
| P(T<=t) two-tail             | 2.38146E-09  |             |
| t Critical two-tail          | 2.20098516   |             |

What each line means:

- **Mean** and **Variance** of each column. The standard deviation is the square root of the variance, or `=STDEV.S(C2:C13)`.
- **Pearson Correlation** between pretest and posttest: learners who scored higher before tended to score higher after.
- **df**, the degrees of freedom: the number of pairs minus one.
- **t Stat**, the t-value. Its sign shows the direction: positive here because the posttest (Variable 1) is higher.
- **P(T<=t) two-tail**, the p-value for a nondirectional hypothesis. 2.38146E-09 means 0.00000000238; report it as p < 0.001.
- **t Critical two-tail**, the critical value for alpha = 0.05 and 11 degrees of freedom.

**5. Decide.** By the toolkit's two approaches: the computed t (17.39) is greater than the critical t (2.201), and the p-value is less than alpha (0.05). Reject the null hypothesis.

**6. Conclude.** The boys' Math scores rose from a mean of 7.00 to 13.58, a difference unlikely to be due to chance (t(11) = 17.39, p < 0.001).

### The girls: a significant drop

Run the same steps on the girls' columns (posttest F14:F25 as Variable 1, pretest E14:E25 as Variable 2):

|                              | Posttest       | Pretest     |
|:-----------------------------|:---------------|:------------|
| Mean                         | 6              | 8           |
| Variance                     | 1.818181818    | 2.727272727 |
| Observations                 | 12             | 12          |
| Pearson Correlation          | 0              |             |
| Hypothesized Mean Difference | 0              |             |
| df                           | 11             |             |
| t Stat                       | -3.249615362   |             |
| P(T<=t) one-tail             | 0.003870259547 |             |
| t Critical one-tail          | 1.795884819    |             |
| P(T<=t) two-tail             | 0.007740519094 |             |
| t Critical two-tail          | 2.20098516     |             |

The t-value is negative and its size, 3.25, exceeds the critical 2.201; the p-value, 0.008, is below 0.05. The null hypothesis is rejected, but in the unexpected direction: the girls' mean fell from 8.00 to 6.00. As the toolkit warns for nondirectional hypotheses, check which group the result favors before calling an intervention effective. Here it would call for reflection: why did the same strategy help the boys and not the girls? The correlation of 0 between the girls' pretest and posttest is also a warning sign that the test or its scoring may not be measuring the same thing twice.

### The same result with one formula

`T.TEST` returns the p-value directly:

`=T.TEST(C2:C13, B2:B13, 2, 1)` returns 2.38146E-09 for the boys, and `=T.TEST(F14:F25, E14:E25, 2, 1)` returns 0.00774 for the girls.

The arguments are the two ranges, the number of tails (2), and the type (1 for paired). Microsoft's documentation lists type 1 as paired, type 2 as "two-sample equal variance (homoscedastic)" and type 3 as "two-sample unequal variance (heteroscedastic)". For a paired test, both ranges must have the same number of scores, or `T.TEST` returns #N/A. To get the t-value itself without the add-in, divide the mean difference by its standard error: in a column D with `=C2-B2` for each learner, `=AVERAGE(D2:D13)/(STDEV.S(D2:D13)/SQRT(COUNT(D2:D13)))`.

## The independent t-test, step by step

The toolkit's third question compares the boys' and girls' posttest scores.

**1. Hypotheses.** Null: there is no difference between the boys and the girls in their Math posttest scores. Research: there is a difference.

**2. Equal or unequal variances?** The boys' posttest variance (6.99) is 3.85 times the girls' (1.82). When one group's variance is several times the other's, the unequal-variances (Welch) test is the safer choice. With equal group sizes, as here, the two versions give the same t-value and very similar p-values, so the decision does not change. The toolkit's answer key uses 22 degrees of freedom, which is the equal-variances test.

**3. Run the test.** Select **Data**, **Data Analysis**, **t-Test: Two-Sample Assuming Equal Variances**. Variable 1 Range: the boys' posttest scores (C2:C13); Variable 2 Range: the girls' (F14:F25); Hypothesized Mean Difference: 0; Alpha: 0.05.

|                              | Boys        | Girls       |
|:-----------------------------|:------------|:------------|
| Mean                         | 13.58333333 | 6           |
| Variance                     | 6.992424242 | 1.818181818 |
| Observations                 | 12          | 12          |
| Pooled Variance              | 4.40530303  |             |
| Hypothesized Mean Difference | 0           |             |
| df                           | 22          |             |
| t Stat                       | 8.850093878 |             |
| P(T<=t) one-tail             | 5.30585E-09 |             |
| t Critical one-tail          | 1.717144374 |             |
| P(T<=t) two-tail             | 1.06117E-08 |             |
| t Critical two-tail          | 2.073873068 |             |

**4. Decide and conclude.** The computed t (8.85) exceeds the critical t (2.074), and p < 0.001. The boys scored higher than the girls on the posttest (t(22) = 8.85, p < 0.001). The Welch test gives the same t, with p < 0.001. The formula `=T.TEST(C2:C13, F14:F25, 2, 2)` returns 1.06117E-08; with type 3, Excel runs the Welch test.

**5. Check the starting point.** A posttest difference only means something if the groups started alike. The boys' and girls' pretest means, 7.00 and 8.00, do not differ (t(22) = −1.32, p = 0.201), so the posttest gap opened during the study. Note that this comparison is causal-comparative: both groups received the same strategy, and the test shows that boys and girls responded differently, not that the strategy caused the difference.

## The toolkit's exercise: the completed answer key

The toolkit's answer key gives the means and standard deviations but leaves the t-values and p-values blank. Here is the table in the toolkit's format, completed:

| Comparison | Mean (SD) | t-computed | df | t-critical (two-tailed, alpha 0.05) | p-value | Decision |
|:--|:--|--:|--:|--:|:--|:--|
| Boys vs girls, posttest | 13.58 (2.64) vs 6.00 (1.35) | 8.85 | 22 | 2.074 | p < 0.001 | Reject the null |
| Boys, pretest vs posttest | 7.00 (2.04) vs 13.58 (2.64) | 17.39 | 11 | 2.201 | p < 0.001 | Reject the null |
| Girls, pretest vs posttest | 8.00 (1.65) vs 6.00 (1.35) | −3.25 | 11 | 2.201 | p = 0.008 | Reject the null; scores fell |

The answer key gives a critical t of 2.10 for all three tests. That value belongs to 18 degrees of freedom, as in the toolkit's earlier boys-and-girls example, where 2.101 is correct. For 11 degrees of freedom the toolkit's own table of critical values gives 2.201; for 22, beyond the end of that table, the two-tailed critical value is 2.074. The decisions do not change, because every t-value is far beyond either cutoff, but use the value that matches your degrees of freedom: Excel's "t Critical two-tail" line gives it.

## Add the effect size

A p-value tells you whether a difference is likely to be more than chance; it does not tell you how big the difference is. With a large class, a tiny gain can be "significant"; with a small class, a big gain may not be. Report an effect size with every test.

- **Paired design (dz):** in column D, `=C2-B2` for each learner; then `=AVERAGE(D2:D13)/STDEV.S(D2:D13)`. For the boys, dz = 5.02; for the girls, −0.94.
- **Two groups (Cohen's d):** the difference between the means divided by the pooled standard deviation: `=(AVERAGE(C2:C13)-AVERAGE(F14:F25))/SQRT(((COUNT(C2:C13)-1)*VAR.S(C2:C13)+(COUNT(F14:F25)-1)*VAR.S(F14:F25))/(COUNT(C2:C13)+COUNT(F14:F25)-2))`. For the posttests, d = 3.61.

Cohen's rough guide calls 0.2 small, 0.5 medium and 0.8 large. The values here are very large because the toolkit's data are made up to teach; real classroom gains are usually much smaller. The paired dz is usually larger than the d for two groups with the same means, because it divides by the spread of the changes, which is smaller when pretest and posttest are correlated; say which one you report.

## A real example: Grade 1 reading in Candon City's schools

The second example uses public data: the end-of-year Comprehensive Rapid Literacy Assessment (CRLA) for SY 2025-26 from the EDCOM II dashboard, and DepEd's selected-school CRLA file, which covers five regions; DepEd does not say which round it covers, but its date and results point to the start of the year. For each of the 22 public schools in the Schools Division of Candon City with at least ten Grade 1 learners assessed in both rounds, the table gives the share of Grade 1 learners who were struggling readers (Low Emerging or High Emerging) in Mother Tongue reading. Schools are coded, not named.

| School    | Grade 1 learners, start   | Struggling at the start (%)   | Grade 1 learners, end   | Struggling at the end (%)   |
|:----------|:--------------------------|:------------------------------|:------------------------|:----------------------------|
| School 1  | 12                        | 100.0                         | 12                      | 0.0                         |
| School 2  | 18                        | 100.0                         | 18                      | 0.0                         |
| School 3  | 25                        | 100.0                         | 25                      | 0.0                         |
| School 4  | 26                        | 100.0                         | 26                      | 7.7                         |
| School 5  | 18                        | 100.0                         | 19                      | 10.5                        |
| School 6  | 48                        | 97.9                          | 49                      | 0.0                         |
| School 7  | 29                        | 93.1                          | 29                      | 6.9                         |
| School 8  | 37                        | 91.9                          | 34                      | 2.9                         |
| School 9  | 18                        | 88.9                          | 17                      | 0.0                         |
| School 10 | 17                        | 88.2                          | 17                      | 35.3                        |
| School 11 | 22                        | 86.4                          | 22                      | 0.0                         |
| School 12 | 25                        | 84.0                          | 25                      | 0.0                         |
| School 13 | 18                        | 83.3                          | 39                      | 15.4                        |
| School 14 | 25                        | 80.0                          | 25                      | 0.0                         |
| School 15 | 15                        | 80.0                          | 15                      | 6.7                         |
| School 16 | 21                        | 71.4                          | 22                      | 9.1                         |
| School 17 | 19                        | 63.2                          | 19                      | 0.0                         |
| School 18 | 98                        | 55.1                          | 100                     | 0.0                         |
| School 19 | 11                        | 54.5                          | 11                      | 0.0                         |
| School 20 | 15                        | 53.3                          | 15                      | 6.7                         |
| School 21 | 44                        | 34.1                          | 44                      | 0.0                         |
| School 22 | 19                        | 31.6                          | 19                      | 0.0                         |

Copy the two percentage columns into Excel and run the paired test with the end-of-year shares as Variable 1:

|                              | End of year   | Start of year   |
|:-----------------------------|:--------------|:----------------|
| Mean                         | 4.598151233   | 78.95319595     |
| Variance                     | 67.37793474   | 457.7513546     |
| Observations                 | 22            | 22              |
| Pearson Correlation          | 0.1848827665  |                 |
| Hypothesized Mean Difference | 0             |                 |
| df                           | 21            |                 |
| t Stat                       | -16.25745895  |                 |
| P(T<=t) one-tail             | 1.12798E-13   |                 |
| t Critical one-tail          | 1.720742903   |                 |
| P(T<=t) two-tail             | 2.25595E-13   |                 |
| t Critical two-tail          | 2.079613845   |                 |

The mean share of struggling readers fell from 79.0% to 4.6%, a drop of 74.4 points, and the share fell in all 22 schools (t(21) = −16.26, p < 0.001, dz = 3.47).

Two lessons follow. First, the unit here is the school, not the learner: each row is one school's share, so the test says that schools' shares fell, not how any child progressed, and the learners assessed may differ between rounds (School 13 assessed 18 Grade 1 learners at the start and 39 at the end). Second, and more important for action research, nothing special was done in these schools: this is what a school year of ordinary teaching does to Grade 1 reading. A teacher who tried a reading strategy in one of these classes and ran a paired t-test would have found a highly significant gain whether or not the strategy helped. [Pretest-Posttest Design and Regression to the Mean](https://nash.com.ph/blog/pretest-posttest-design-regression-to-the-mean) explains why, and what to do about it.

## How to report the results

In the text: *The boys' Math scores rose from a mean of 7.00 (SD 2.04) on the pretest to 13.58 (SD 2.64) on the posttest, t(11) = 17.39, p < 0.001, dz = 5.02.*

In a table, the toolkit's format works well: group, mean, SD, t-computed, t-critical and p-value, with the degrees of freedom and the effect size added. Put the full Excel output in an appendix. [DepEd Action Research Report Format, With a Sample](https://nash.com.ph/blog/action-research-report-format) shows where results go in the report.

## Common mistakes

- **Using the two-sample test for a pretest and posttest of the same class.** The same learners measured twice need the paired test; the two-sample test ignores the pairing and usually understates the change.
- **Reading the one-tail p-value.** Use the two-tail line unless you wrote a directional hypothesis before collecting data.
- **Ignoring the sign.** A significant negative t means scores went down.
- **Treating "significant" as "effective".** A one-group gain cannot separate the intervention from ordinary growth; use a comparison group.
- **Testing many things and reporting the one that worked.** Decide the tests before you see the data, and report all of them.
- **Leaving out learners without saying so.** Report how many learners had both scores, and why others were dropped.
- **Using a critical value that does not match your degrees of freedom.** Use Excel's printed value or the table row for your df.

## When the t-test is not enough

The t-test assumes that the scores, or the differences between pretest and posttest, are roughly normally distributed, which matters most in very small groups. With fewer than about 15 learners and skewed scores, or with ranked data such as reading levels, a rank-based test such as the Wilcoxon signed-rank test (paired) or the Mann–Whitney test (two groups) is safer; Excel has no built-in tool for them, but [Wilcoxon and Mann-Whitney Tests in Excel for Small Classes](https://nash.com.ph/blog/wilcoxon-mann-whitney-test-excel-action-research) shows how to compute them by hand, and free programs such as JASP or jamovi do them directly. For three or more groups, the toolkit points to analysis of variance (ANOVA), which is also in the Analysis ToolPak, and for counts or categories, the chi-square test.

## Frequently asked questions

### What is the difference between a paired and an independent t-test?

A paired t-test compares two measurements of the same learners, such as a pretest and a posttest. An independent t-test compares two different groups, such as an experimental and a control section. In Excel, use "t-Test: Paired Two Sample for Means" for the first and "t-Test: Two-Sample Assuming Equal Variances" or "Unequal Variances" for the second.

### How do I do a t-test in Excel without the Data Analysis tool?

Use the `T.TEST` function: `=T.TEST(range1, range2, tails, type)`, with tails 2 for a nondirectional hypothesis and type 1 for paired, 2 for two groups with equal variances, or 3 for unequal variances. It returns the p-value. It also works in Google Sheets.

### Which p-value do I use in the Excel output?

"P(T<=t) two-tail", unless you stated a directional hypothesis (for example, "the posttest mean will be higher") before collecting the data. Reject the null hypothesis if it is below your alpha, usually 0.05.

### What are the answers to the t-test exercise in the DepEd Action Research Toolkit?

Boys versus girls on the posttest: t(22) = 8.85, p < 0.001. Boys' pretest versus posttest: t(11) = 17.39, p < 0.001. Girls' pretest versus posttest: t(11) = −3.25, p = 0.008, a drop. All three null hypotheses are rejected. The two-tailed critical values are 2.074 for 22 degrees of freedom and 2.201 for 11.

### Is a significant t-test proof that my intervention worked?

No. A paired t-test on one class shows that scores changed, not why. Learners grow over a school year, and low scorers tend to rise on a retest, so a comparison group and an effect size are needed before you credit the intervention.

## More on DepEd research

- Start here: [DepEd Action Research Guide: From Topic to Report](https://nash.com.ph/blog/deped-action-research-guide)
- [Sample Action Research in DepEd Format: A Complete Example](https://nash.com.ph/blog/sample-action-research-deped)
- [Pretest-Posttest Design and Regression to the Mean](https://nash.com.ph/blog/pretest-posttest-design-regression-to-the-mean)
- [One-Way ANOVA in Excel for Action Research](https://nash.com.ph/blog/anova-excel-action-research)
- [Action Research Sample Size: Slovin, Power and Sampling](https://nash.com.ph/blog/action-research-sample-size-sampling)
- [DepEd Action Research Proposal Format, With a Sample](https://nash.com.ph/blog/action-research-proposal-format)
- [DepEd Action Research Report Format, With a Sample](https://nash.com.ph/blog/action-research-report-format)
- [DepEd Research Management Guidelines: DO 16, s. 2017](https://nash.com.ph/blog/deped-research-management-guidelines)
- [Action Research and Thesis Topics on Reading (CRLA)](https://nash.com.ph/blog/action-research-topics-reading-crla)
- Related: [CRLA Start vs End of SY 2025-26 in the Same Schools](https://nash.com.ph/blog/crla-start-vs-end-of-year-sy-2025-2026), [How to Read a School's CRLA Results Fairly](https://nash.com.ph/blog/how-to-read-school-crla-results)

## Sources

- Department of Education and the Basic Education Sector Transformation (BEST) Program. (2019). *Action Research Toolkit: Teacher's Guide*, sections 3.1 (experimental designs) and 5.3 (inferential statistics), with support from the Australian Government. The practice data and answer key are from section 5.3.
- Microsoft. [T.TEST function](https://support.microsoft.com/en-us/excel/t-test-function) and [Load the Analysis ToolPak in Excel](https://support.microsoft.com/en-us/Excel/load-the-analysis-toolpak-in-excel).
- Cohen, J. (1988). *Statistical power analysis for the behavioral sciences* (second edition). Lawrence Erlbaum Associates.
- CRLA results: the [EDCOM II early-childhood dashboard data](https://nash.com.ph/edcom-data), cleaned by NASH Reviewer, end of SY 2025-26 as of 10 May 2026, and DepEd's start-of-year CRLA file for SY 2025-26; see [CRLA Start vs End of SY 2025-26 in the Same Schools](https://nash.com.ph/blog/crla-start-vs-end-of-year-sy-2025-2026).

*Analysis and chart by Mark Anthony Llego for NASH Reviewer. NASH Reviewer is not affiliated with DepEd or EDCOM II.*

---

© 2026 Mark Anthony Llego, NASH Reviewer. All rights reserved. You may quote up to 100 words from this article with the author's name, its title and a link. In research, cite it as: Llego, M. A. (2026, October 28). *Paired and Independent t-Test in Excel for Action Research*. NASH Reviewer. https://nash.com.ph/blog/t-test-excel-action-research. Any other reuse of the article, its charts or its data needs written permission: email support@nash.com.ph. Terms of Service: https://nash.com.ph/terms#website-content-and-data