Explore chapters and articles related to this topic
What are you measuring in your research?
Published in Perry R. Hinton, Isabella McMurray, Presenting Your Data with SPSS Explained, 2017
Perry R. Hinton, Isabella McMurray
The results of a research study have to be input into SPSS in the form of a spreadsheet. This means that all the results have to be input as numbers or numeric codes. As we have seen, this is fine for scale data as these responses are already in numeric form – they are measurements along a scale. However, with nominal or ordinal data (categorical data), the responses have to be given number codes that are then input into SPSS, such as the number 1 for a tick in the ‘female’ box and a number 2 for a tick in the ‘male’ box for a question on gender. It is a good idea to document the particular choice of number codes so that you do not make a mistake in inputting your results. Also, if the codes are all written down then more than one person can input the data (using the designated codes) without making errors – as everyone is using the same codes. This document is often referred to as a codebook (even though it might simply be a single sheet of paper). You might already have undertaken surveys, either using paper-and-pen or completed online and never heard of the word ‘codebook’ before. If you have undertaken a very short paper-and-pen survey you may have added up the different responses without bothering to make a codebook. Alternatively, online survey programs automatically create a codebook for you – but do not specifically tell you that they are doing it. A ‘codebook’ is not an actual book, in the same way an Excel ‘workbook’ is not a book. A codebook is a summary table of all of your outcome measures and the decisions you have made as to what to name them, what types of numbers you are collecting and the numeric codes and written labels you have chosen for them.
Tidy data
Published in Benjamin S. Baumer, Daniel T. Kaplan, Nicholas J. Horton, Modern Data Science with R, 2021
Benjamin S. Baumer, Daniel T. Kaplan, Nicholas J. Horton
The codebook is a document—separate from the data table—that describes various aspects of how the data were collected, what the variables mean and what the different levels of categorical variables refer to. The word codebook comes from the days when data was encoded for the computer in ways that make it hard for a human to read. A codebook should include information about how the data were collected and what constitutes a case. Figure 6.3 shows the codebook for the HELPrct data in the mosaicData package. In R, codebooks for data tables in packages are available from the help() function.
Tidy data and iteration
Published in Benjamin S. Baumer, Daniel T. Kaplan, Nicholas J. Horton, Texts in Statistical Science, 2017
Benjamin S. Baumer, Daniel T. Kaplan, Nicholas J. Horton
The codebook is a document—separate from the data table—that describes various aspects of how the data were collected, what the variables mean and what the different levels of categorical variables refer to. The word codebook comes from the days when data was encoded for the computer in ways that make it hard for a human to read. A codebook should include information about how the data were collected and what constitutes a case. Figure 5.2 shows the codebook for the babynames data in Table 5.1. In R, codebooks for data tables are available from the help() function.
Bridging the Trust Gap in Influencer Marketing: Ways to Sustain Consumers’ Trust and Assuage Their Distrust in the Social Media Influencer Landscape
Published in International Journal of Human–Computer Interaction, 2022
Chung-Wha (Chloe) Ki, Tsz Ching Chow, Chunsheng Li
The focus group data were audio-recorded with the participants’ permission and then transcribed into written form for analysis. The data analysis followed the open coding process Corbin and Strauss (1990) proposed. Open coding refers to segmenting qualitative data into meaningful expressions and describing those expressions according to relevant concepts or themes (Asan et al., 2017; Jain & Roy, 2016). First, the focus group discussion leader assessed the data and developed a preliminary code list. The code list was shared with the moderator to finalize the list and create a codebook. While referring to the codebook, both researchers began to code the data, which is the process of transforming the textual information into a set of meaningful and cohesive concepts. After the coding was completed, a third judge who did not participate in the coding process verified all entries in the concepts identified. An agreement percentage was then computed to quantify the intercoder reliability. The reliabilities were suitable, as they ranged from 90 to 95%, which is above the 85% threshold recommended (Kassarjian, 1977).
A vision-based fall detection framework for the elderly in a room environment using motion features and DAG-SVM
Published in International Journal of Computers and Applications, 2022
Huazheng Zhu, Jinglong Du, LuLu Wang, Baoru Han, Yuanyuan Jia
In this section, popular K-means algorithm is applied to analyze OMFVs. For training process, the OMFVs are clustered by K-means algorithm to generate a codebook which is used as index mapping table of OMFVs. Meanwhile, we record the center and label of each cluster, and each code in the codebook denote one cluster. Then we count the frequency of each code that occurs in a video sequence by Term Frequency (TF) mode [32,33] which is widely used in text information retrieval technology. And the frequency can be calculated as: where and is the frequency and the number of times that code c occurs in video sequence s, respectively. The denominator is the number of times that all codes occur in video sequence s with k = 1, 2, …, K. Finally, these frequencies generate the input vector of classifier according to the order of the code in the codebook. After this step, the feature vectors of dimensions are reduced to K dimensions. Then the obtained input feature vectors are employed to train a classifier for distinguishing different activities.
A novel image compression technology based on vector quantisation and linear regression prediction
Published in Connection Science, 2021
Shuying Xu, Chin-Chen Chang, Yanjun Liu
Before explaining the pixel prediction by linear regression, preliminaries will be introduced first. Since the original image will be divided into blocks, a codebook containing codewords with the size of can be obtained by the LBG algorithm. The core idea of LBG algorithm is to approximate the optimal regenerated codebook by training vector sets and some iterative algorithms. First, we select several typical images and divide them into blocks as the training block set; secondly, we select blocks as code vectors randomly; and then, we divide all the blocks according to the code vectors so that each block in the set has the smallest distance from the corresponding code vector. After that, the centroids of sets are calculated to obtain new code vectors. Finally, we conduct multiple iterations until the code vector converges. Thus, the codebook is consisting of the latest K code vectors.