Likert Scale Analyzer
Paste Likert-scale responses — one row per respondent, one column per item — and get every item's median, mode, mean, SD and percentage agreeing, a diverging stacked bar chart you can download, and a composite score for each respondent with negatively worded items reverse-coded. It runs in your browser and nothing is uploaded. Free, no sign-up.
Paste from Excel, Google Sheets or SPSS, or type numbers separated by commas or spaces. Blank cells and NA count as missing; option labels such as Agree are read as their codes.
Results
Paste at least one respondent's answers, or load the example, to see item statistics.
How to Use the Likert Scale Analyzer
- 1. Paste your responses. One row per respondent and one column per item, straight from Excel, Google Sheets or SPSS. A header row and an ID column are detected for you — untick either box if the guess is wrong.
- 2. Describe the scale. Choose 4, 5 or 7 points, say whether codes start at 1 or 0, and edit the option labels. A cell holding a label such as "Agree" is read as that option, which is how Google Forms exports answers.
- 3. Read the item table. Each item gets n, median, mode, mean, SD and the share disagreeing, neutral and agreeing — downloadable as CSV.
- 4. Export the chart. Save the diverging stacked bar chart as SVG for a paper or PNG for slides.
- 5. Score the scale. Tick the negatively worded items to reverse-code them, choose sum or mean, and download one composite score per respondent.
What Each Statistic Tells You About a Likert Item
A Likert item is one statement with ordered answer options; a Likert scale is several items about the same idea, combined into one score (Likert, 1932). The options are ranked, but nothing guarantees that the step from Disagree to Neutral is the same size as the step from Neutral to Agree. That is why the analyzer reports two kinds of summary side by side.
| Statistic | What it describes | Assumes equal spacing? |
|---|---|---|
| Median | The middle answer once all answers are put in order | No |
| Mode | The most common answer — every tied option is listed | No |
| % disagree / neutral / agree | The share of answers below, at and above the midpoint | No |
| Top 2 (7-point scales) | The share choosing one of the two highest options | No |
| Mean and SD | The average code and its spread (SD divides by n − 1) | Yes |
On 4- and 5-point scales, % agree already is the top-two-box figure, so the Top 2 column appears only for 7-point scales.
For a single item, lead with the median and the percentages. Jamieson (2004) argued that means of ordinal answers can mislead; Norman (2010) showed that the usual parametric tests are robust on Likert data, especially for scale scores built from several items. Reporting both, as the table does, shows the distribution as well as the average.
Worked Example: A Workshop Evaluation
Press Load example. Twelve teachers rated a professional-development workshop on five 5-point statements (1 = Strongly disagree, 5 = Strongly agree). Q3, The pace was too fast, is negatively worded, and teacher T03 left Q5 blank. Three rows of the item table read:
| Item | n | Median | Mode | Mean | SD | Disagree | Neutral | Agree |
|---|---|---|---|---|---|---|---|---|
| Q1 | 12 | 4 | 4 | 3.92 | 0.90 | 8.3% | 16.7% | 75% |
| Q3 | 12 | 2 | 2 | 2.25 | 1.06 | 66.7% | 16.7% | 16.7% |
| Q5 | 11 | 3 | 3, 4 | 3.36 | 0.92 | 18.2% | 36.4% | 45.5% |
Q5's n is 11 because the blank is skipped, not counted as a zero, and it has two modes because 3 and 4 were each chosen four times. Q3 looks like the weakest item until you remember its wording: two-thirds of teachers disagreed that the pace was too fast, which is good news. On the chart, Q3's bar sits mostly left of the center line while the other four lean right — the diverging layout makes a reversed item easy to spot before you score anything.
Reverse Coding and Composite Scores
Before summing, a negatively worded item has to point the same way as the rest. Reverse-coding replaces each answer with (lowest code + highest code) − answer, so on a 1–5 scale a 2 becomes 4 and a 3 stays 3. Tick Q3 in the composite section and T01's answers 4, 5, 2, 4, 3 become a sum of 4 + 5 + 4 + 4 + 3 = 20 instead of 18.
With Q3 reversed, the sum is given to the 11 teachers who answered all five items: mean 18.64, SD 4.48, range 11–24. T03 gets no sum, because a total over four items is not comparable with a total over five. Switch to Mean and every teacher is scored on the items they answered — T03's mean is 4 over four items — giving a mean of 3.75 and an SD of 0.86 across all twelve. A mean keeps the original 1–5 metric, which is often easier to explain; a sum matches how many published scales are scored.
Check reliability before you add items together
A composite only makes sense if its items measure the same thing. Paste the same data into the Cronbach's Alpha Calculator, with the same items reversed, before you report a total.
Reading the Diverging Stacked Bar Chart
Each bar is one item, split into one segment per option. Disagreement runs left of the center line, agreement right, and the neutral option is split in half across the line, so the two ends of a bar show the full share leaning each way — the layout Heiberger and Robbins (2014) recommend for comparing items at a glance. Percentages are printed inside segments wide enough to hold them, hovering shows a segment's exact share and count, and the Okabe–Ito vermillion-to-blue colors stay distinguishable with the common color-vision deficiencies. The chart is drawn on white with its own fonts, so the downloads match the screen.
Frequently Asked Questions
- Can you calculate a mean for Likert scale data?
- You can, and most published research does, but it treats the codes as equally spaced numbers. For a single item, the median and the percentage agreeing are the safer headline, with the mean as a secondary figure. For a composite of several items, the summed or averaged score behaves much more like an interval measure and means are standard practice.
- What is the difference between a Likert item and a Likert scale?
- A Likert item is one statement with ordered answer options, such as Strongly disagree to Strongly agree. A Likert scale is a set of items measuring the same construct, combined into one score by summing or averaging — which is why composite scores, not single items, are the natural unit of analysis.
- How do I calculate percent agree for a Likert item?
- Divide the number of respondents who chose an option above the midpoint by the number who answered the item, ignoring blanks. On a 5-point scale that is Agree plus Strongly agree, often called top-two-box. On a 7-point scale the analyzer reports both: % agree counts all three options above the midpoint, and Top 2 counts only the two highest.
- What is a diverging stacked bar chart?
- A horizontal stacked bar chart aligned on the scale's midpoint rather than its left edge: negative answers extend left, positive answers extend right, and the neutral option straddles the center line. It is the standard way to compare several Likert items at once.
- How should I handle missing Likert responses?
- The analyzer skips blanks item by item, so each item's n is the number who answered it. A composite sum is only computed for complete rows, while a mean uses the items each person answered. Scale manuals often set a minimum number of answered items for a person-mean; follow yours and report how many respondents were left out.
- Which statistical test should I use for Likert data?
- For a single item, compare groups with a rank-based test such as Mann–Whitney U for two independent groups or Wilcoxon signed-rank for paired answers. For composite scale scores, t-tests and ANOVA are widely used and generally robust. Whatever you choose, state it and report the medians or percentages alongside any means.