Likert Scales - Reverse Scored/Negative Items (Part 1)
HTML-код
- Опубликовано: 17 июн 2014
- This video covers how to reverse code negatively worded Likert items in SPSS. Learn how to recode all negative Likert items both efficiently and accurately in SPSS. Lots more Likert & SPSS Videos here: www.udemy.com/survey-data/?co...
RUclips Channel (Quantitative Specialists): / statisticsinstructor
Subscribe today!
Video Transcript: In this video we'll take a look at how to deal with reverse scored items on a Likert scale. And these are also known as negative items. For our example here we're looking at an example where we're assessing perfectionism, and we're using the Multidimensional Perfectionism Scale and you can see the reference here for those of you who may be interested in looking this up and pulling the article. Now what I did in this example was I just selected a few of the questions from the Self-Oriented Perfectionism subscale of this scale here. So we've got four questions here, and these are also called items, so I'll be referring to these items often as well. And a person reads each question or item, such as the first one, "One of my goals is to be perfect and everything I do." And if you think of someone here who is highly perfectionistic, or we think of them as a perfectionist, what is it that they would answer to this question? Well, notice the scale ranges from strongly disagree to strongly agree, and we have numbers here to correspond to that continuum. So if someone's highly perfectionistic and they read this question, "One of my goals is to be perfect in everything I do," they should answer with either a 6 or 7, right? Probably a 7, but a 6 is possible as well. But whatever their answer may be, it definitely should be on that high end of this continuum. They should indicate to some degree that they do strongly agree with the statement. Our next question, let's go ahead and read that, "I strive to be as perfect as I can be." Once again, if we're thinking of a perfectionist, they should answer with a 6 or 7 again. Now the next question is what I want to focus on here for a minute, really these last two questions. For question number three, "I never aim for perfection on my work." Now if we think about someone there who is a perfectionist, and they read this question "I never aim for perfection on my work," what should they do there, you want to ask yourself? Well they most likely would strongly disagree, right? With a 1, answer with a 1, or maybe answer with a 2. And then as we read this last question, "I seldom feel the need to be perfect," well a perfectionist definitely doesn't feel that way. They feel the need to be perfect usually all the time, so they're going to answer, again, on the strongly disagree end, so with a 1 or 2. These last two questions or items, where someone who is high on the trait or we call it a construct being measured here, perfectionism, someone who is high on that variable, where they answer low on the scale, those are known as negative or reverse scored items. So these are special, because we have to do something with these prior to creating a total score. And what we'll often do here is, we'll add up these, the scores on these items here, to create a total score. So, for example, if someone answered all 4s, then they would get, if the scale consisted of just four questions, they would get a 16, 4 + 4 + 4 + 4, and so on. Now in another video I do go through how to add up items on a number of different questions, or put another way, how to add up a number of different variables, so you may want to take a look at that, if you'd like more information on how to do that. But in this video, we're going to deal with just how do I handle these reverse coded items, how do I use SPSS to actually change the responses on a scale? So let's go ahead and look at that. So once again a perfectionist answers with 7s on one and two and 1s on 3 & 4, or somewhere close to that. And the problem that happens here is that the instructions from the authors or the creators of the scale, as it's usually the case, is that high scores on the scale, since it's called perfectionism, high scores indicate someone who is highly perfectionistic. But if we look at this here, if we take the person's responses, let's say we have someone to respond in the most extreme way, so they got 7s on questions one and two, and they got 1s on questions three and four. Well then their score would be, if we just literally added these up, it would be 7 + 7 + 1 + 1, which will give us 16. Now if you think about that, that is not a high score, because the highest possible score according to the scale is 28, or seven times four items. So something's wrong here. We need to work with this scale, we need to do something in order to get it so that someone who is high on perfectionism, doesn't score 16, but instead they score something like a 28. So what we need to do in SPSS is we need to do is called reverse coding these items here.
Really helpful. Thank you
This is pretty helpful, thank you very much. But I am still confused about how to analyse data. After I coded likert scale in SPSS, each variables might includes 3 or 4 items, like in your this vedio, in terms of self-orientation perfectionism, there are 4 items to measure self-orientation perfectionism, I have no idea how to analyse the correlation between two variables. For example, Q7-1, Q7-2, Q7-3 measure 'employee motivation' , Q8-1,Q8-2, Q8-2 measure 'employee performance', how I can analyse the correlation?
For a measure with a scale >1, you can recode by taking the maximum value, adding 1, and then subtracting the original score. The result will be a reverse coded scale.
Thank you so much, I was struggling to code a vector command in Maltab that will transform a "negatively" ranked result to a positive one. Can you please explain the logic behind it too?
Your trick actually eased out my tension. I was looking for this since a long time
@@tomyishay7745 take the original value and subtract twice the original value. It's al algabraic function that takes the original value to 0 and then continues in the same direction by the absolute value of the original number.
Hallo there, If I report the item wording and compare on the Item level, do I need to recode my reversed scored items? For running a MANOVA with negative and positive items do I need to recode my negative items? Or: do I only recode when I want to analyse the factorial level? Thank you for your help!
So, if you are not doing any regression or factor analysis, just Mann Whitney you wouldn’t necessarily have to reverse code, just interpret results correctly? Is that correct?
i also have problems in coding as i have to find relationship between two variables.so if i do corelation i put independent variables in the set but only one factor is possible to enter as i have many question item in the independent list as well..how to do this plzzz guide
I have a seperate section on DIFFICULTIES (section 02 of my questionnaire), where the statements are positive e.g., I have difficulty using PC) but about difficulties. Should I reverse code the items?
What's the posh name for doing this?
Is it meant that we should frame all questions either positive or negative???please answer
does anyone know if there is reverse scoring in the Timmermans Tinder Motivation Scale. i cant figure it out x
❤
How can I interpret 1: strongly disagree, 2: disagree, 3: somewhat disagree, 4: somewhat agree, 5: agree, 6: strongly agree please? I didn't find any reference
Good video but you afford to talk through it quicker. People who are searching for this video aren't looking for a full explanation of everything.
Typical of amateur or new youtube video - the need to over explain. Some one looking for reverse scoring is probably familiar with a Likert scale. And it is frustrating to go through four minutes of background knowledge most viewers already know to get to the minute of information of interest.
I don't know why content creators fall into this practice. I think it's an inability to identify their target audiences. The uploader may not be sure who he is creating content for and therefore assumes he has to explain everything to everyone. However, it is a reasonable assumption that a person already familiar with rudimentary aspects of data collection and analyses would be searching for something specific and relatively more advanced. In conclusion, I would argue the content count be tailored to the needs of its specific audience thus making it more valuable.
I'm really satisfied with the speed of speech since i'm not a fast learner!!!>
I'd rather him go through it properly than quickly and me having questions later when its time to actually create a scale IRL. He's a good teacher.