**Chronbach’s Alpha** is a way to measure the internal consistency of a questionnaire or survey.

Cronbach’s Alpha ranges between 0 and 1, with higher values indicating that the survey or questionnaire is more reliable.

When reporting the value of Cronbach’s Alpha in a final report, you need to include the following two values:

- The number of items used on the subscale.
- The value of Cronbach’s Alpha.

The following examples show how to report Cronbach’s Alpha in different situations.

**Example 1: Reporting Cronbach’s Alpha for One Subscale**

Suppose a restaurant manager wants to measure overall satisfaction among customers. She decides to send out a survey to 200 customers who can rate the restaurant on a scale of 1 to 5 for 12 different categories.

When she receives the results of the survey, she finds that the value for Cronbach’s Alpha is 0.84.

Here is how she would report Cronbach’s Alpha in a formal write-up:

A satisfaction survey was sent to 200 customers. The survey consisted of 12 items and the value for Cronbach’s Alpha for the survey was α = .84.

**Example 2: Reporting Cronbach’s Alpha for Multiple Subscales**

Suppose a Human Resources manager at a company sends out a three-part questionnaire to all 500 employees at the company.

When she receives the results of the survey, she calculates the value for Cronbach’s Alpha for all three of the subscales.

Here is how she would report the value of Cronbach’s Alpha in a formal write-up:

A three-part questionnaire was sent to 500 employees. The agreeableness subscale consisted of 10 items (α = .65), the leadership subscale consisted of 12 items (α = .82), and the overall satisfaction subscale consisted of 14 items (α = .88).

**Notes**

The following table describes how different values of Cronbach’s Alpha are usually interpreted:

Cronbach’s Alpha |
Internal consistency |
---|---|

0.9 ≤ α | Excellent |

0.8 ≤ α < 0.9 | Good |

0.7 ≤ α < 0.8 | Acceptable |

0.6 ≤ α < 0.7 | Questionable |

0.5 ≤ α < 0.6 | Poor |

α < 0.5 | Unacceptable |

When reporting the value of Cronbach’s Alpha for a given scale or survey, you can reference this table to determine if the value is at least considered “acceptable” or not.

**Additional Resources**

Cronbach’s Alpha Calculator

How to Calculate Cronbach’s Alpha in Excel

How to Calculate Cronbach’s Alpha in R

How to Calculate Cronbach’s Alpha in Python

Is it possible to calculate the overall Cronbach’s alpha if you are given the alpha values for each of the subscales? If so, how would you calculate it?

Excellent overview of Cronbach’s alpha – thanks for sharing it.

You might want to make a minor edit to correct the spelling in the first paragraph where there is an “h” between C and r in the word.

Can you share a reference for the interpretation?

if one of the questionnaires has an item that lowers the Cronbach’s alpha regularly(0.643) and even lowers it when reversed (0.485)

what would be the optimal decision to make in accordance to the other statistical calculations

Hi Ameer…When an item in a questionnaire consistently lowers the Cronbach’s alpha, both in its original and reversed form, it indicates that the item may not be measuring the same underlying construct as the other items. Here are the steps and considerations for making an optimal decision based on this situation:

### Steps to Address the Problematic Item

1. **Examine the Item’s Content**:

– Review the wording and content of the item to ensure it aligns with the construct being measured. Sometimes ambiguous or poorly worded items can cause issues.

2. **Check for Consistency with Other Items**:

– Evaluate how well the item correlates with other items. Low or negative correlations with most items suggest the item is not consistent with the scale.

3. **Analyze Item-Total Correlation**:

– Calculate the item-total correlation for the problematic item. A low item-total correlation indicates the item does not correlate well with the overall scale.

4. **Consider Removing the Item**:

– If the item consistently lowers the Cronbach’s alpha and does not align with the construct, consider removing it from the scale. This should ideally result in an increase in the overall reliability.

### Statistical Calculations and Decision Making

1. **Cronbach’s Alpha**:

– Calculate Cronbach’s alpha with and without the problematic item. Compare the values to see the impact of removing the item.

– If the alpha increases significantly upon removal, this supports the decision to remove the item.

2. **Item-Total Correlation**:

– Examine the item-total correlation for each item. Items with very low or negative correlations should be scrutinized.

– Typically, items with an item-total correlation below 0.3 are considered for removal.

3. **Inter-Item Correlation Matrix**:

– Review the inter-item correlation matrix to see how the problematic item correlates with each other item.

– If the problematic item has low correlations with most other items, this is another indication it does not belong.

4. **Factor Analysis**:

– Conduct an exploratory factor analysis (EFA) to see if the problematic item loads onto the same factor as the other items.

– Items that do not load well onto any factor or load onto a different factor may need to be revised or removed.

### Example Calculation

Assume the original Cronbach’s alpha with all items is 0.7. Removing the problematic item increases Cronbach’s alpha to 0.8. This suggests the item is not contributing positively to the reliability of the scale.

– **Original Cronbach’s Alpha**: 0.7

– **Cronbach’s Alpha without Problematic Item**: 0.8

– **Item-Total Correlation of Problematic Item**: 0.1

– **Inter-Item Correlation with Problematic Item**: Mostly below 0.3

### Decision

Given the statistical evidence, the optimal decision is likely to **remove the problematic item**. This decision is based on:

– The increase in Cronbach’s alpha upon removal.

– The low item-total correlation.

– The low inter-item correlations.

– Consistency of these findings with the content review and possibly factor analysis.

By removing the item, you ensure the remaining items on the questionnaire are more consistent and reliable in measuring the underlying construct. This leads to a more valid and reliable assessment tool.