Hi there,
We have sent out a survey using a likert based measuring tool and have now closed the survey.
Whilst inputing the data for analyses, we have realised that there has been a whole question (1 item) missed in the survey. As such, we are unsure of our options. I have tried to search for information on what to do in such a case to ensure that what we are measuring is still the same with the missing item, for comparison reasons.
The item is 1 of 20 that is measuring one variable.
Does anyone have any advice on this, or any articles we can refer to regarding such a matter?
Thanking you in advance.
Concerning the missed data items in general, the researcher must have a strategy to handle with such cases. In you case, the data is related to a complete variable, therefore I suggest one of the following:
1- Repeat data collection.
or
2- Omit the variable if possible
Thank you
Delete the item if not substantively detrimental to your construct measure (recommendable over imputation or editing approaches).
Hope this helps,
Matthew
Actually, you have two options. One is called the HALF-SCALE RULE (HSR) and the other is FULL MISSING SCORE ESTIMATION (FMSE) method.
In HSR, "a score can be calculated if the respondent answers at least 50% of the items in a multi-item scale. In such cases, the recommended algorithm substitutes an estimate for the missing item data that is based on the respondent’s answers to the other items." Here, the missing response is estimated to be the average of the other related responses.
In FMSE, "the missing item responses in a given scale are assumed to be the same as the response to the scale’s answered item and the final item response values are then assigned accordingly."
Source: Maruish, M. E. (Ed.). User’s manual for the SF-36v2 Health Survey (3rd ed.). Lincoln, RI: QualityMetric Incorporated. 2011.
NB: These are however peculiar for the SF-36 tools, but I believe you can adopt either of the methods if your options are limited.
- Badges
- Science topic