HLSS500 Data Synthesis Interpretation and Presentation Discussion Paper Please post an abstract that accurately reflects your study up to this point. If yo

HLSS500 Data Synthesis Interpretation and Presentation Discussion Paper Please post an abstract that accurately reflects your study up to this point. If you need some examples of abstracts to help you get started take a look at some of the abstracts found within the peer-reviewed journals you referenced within your paper. Also, what has been the hardest section to write for this project? What section was the easiest for you to write? How might you work to apply what you learned in this course moving forward across your program of study?Instructions: Fully utilize the materials that have been provided to you in order to support your response. Your initial post should be at least 500 words. Descriptive Statistics
In: Research Training for Social Scientists
By: Vernon Gayle
Pub. Date: 2011
Access Date: May 20, 2019
Publishing Company: SAGE Publications Ltd
City: London
Print ISBN: 9780761963516
Online ISBN: 9780857028051
DOI: https://dx.doi.org/10.4135/9780857028051
Print pages: 363-384
© 2000 SAGE Publications Ltd All Rights Reserved.
This PDF has been generated from SAGE Research Methods. Please note that the pagination of the
online version will vary from the pagination of the print book.
SAGE
SAGE Research Methods
2000 SAGE Publications, Ltd. All Rights Reserved.
Descriptive Statistics
VernonGayle
Statistics
After the initial panic and terror which many social scientists experience when they hear the word statistics,
one of three definitions is conjured up. The original meaning of the word is ‘State-istics’, facts and figures
which are collected by and for the state. In modern times statistics have become a branch of ‘applied
mathematics’, and most British universities offer statistics degrees. The third definition is a ‘statistical test’.
In terms of data analysis, as social scientists, we are interested in applying statistical tests to social science
data.
In social science data analysis there are two types of statistics in which we are interested. The first category
are known as descriptive statistics and the second category are known as inferential statistics. As the name
suggests, descriptive statistics describe something such as a characteristic of the sample. In a general sense
we are used to descriptive statistics, such as a percentage of people who display a certain behaviour, or the
average income of a particular social group, and so on. Inferential statistics, as the name suggests, allow
us to infer or make some inference about an aspect of the social world. Most social research projects use a
mixture of descriptive and inferential statistics.
Statistics that are concerned with only one variable at a time, for example, age or gender are known as
univariate statistics. Statistics that are concerned with two variables, for example, the relationship between
gender and levels of education are known as bivariate statistics. Statistics which are concerned with more
than two variables, for example, the relationship between gender and ethnicity and levels of education are
known as multivariate statistics.
A variable that explains an outcome is known as an explanatory or independent variable. It is typically denoted
by the symbol X. A variable which measures an outcome is known as a dependent variable and is typically
denoted by the symbol Y.
The Data
In the next two chapters the examples and the analysis is undertaken from data extracted from the Youth
Cohort Study of England and Wales (YCS). This is a major programme of longitudinal research designed to
Page 2 of 30
Research Training for Social Scientists
SAGE
SAGE Research Methods
2000 SAGE Publications, Ltd. All Rights Reserved.
monitor the behaviour and decisions of representative samples of young people aged 16 to 19 as they reach
minimum school leaving age and either stay on in education or enter the labour market. The survey collects
information on the young people’s experiences of education, training and work as well as information on their
aspirations, their family and their personal circumstances. Cohort members are contacted by post three times.
The three sweeps of data collection are undertaken at yearly intervals, when the young people are 16–17
(sweep 1), 17–18 (sweep 2) and 18–19 (sweep 3).
The Data Files
The data files can be downloaded from the web http://www.stir.ac.uk/appsocsci/vernon/datafiles2.htm. (There
are a full set of instructions on this web page.)
GETTING STARTED IN SPSS
These instructions assume that you know a little about computers and operating in a Windows environment.
To open a file in SPSS see Figure 25.1.
Page 3 of 30
Research Training for Social Scientists
SAGE
SAGE Research Methods
2000 SAGE Publications, Ltd. All Rights Reserved.
FIGURE 25.1 Starting SPSS
Most people have their own profiles set up in Windows 95 so their screens all look different. To open a data
file go to the Start buttom at the bottom left of your screen. Choose SPSS from the relevant menu for your
set-up (this is sometimes in Data Analysis).
Choose Open an exiting file.
Select the required file by moving the pointer onto a file (e.g. ycs.sav) and click on it. Then click on Open.
The frontmost window in which a grid is displayed is the ‘Data Editor’ window (Figure 25.4). This is where
data are entered or changed, and it is the window you will be using most during the rest of this exercise. Each
column of the grid represents a variable (typically a question) and each row, a case (often a respondent). The
second window, entitled ‘Output1’, is used by SPSS for displaying the results of statistical analyses (see, for
example, Figure 25.10). The large window in the background is called the ‘Application’ window (or sometimes
the ‘Main’ window) and its main use is ‘to hold everything together’. To toggle (switch) between windows go
to Window and click on the window you require.
Page 4 of 30
Research Training for Social Scientists
SAGE
SAGE Research Methods
2000 SAGE Publications, Ltd. All Rights Reserved.
FIGURE 25.2 Opening an existing file
Page 5 of 30
Research Training for Social Scientists
SAGE
SAGE Research Methods
2000 SAGE Publications, Ltd. All Rights Reserved.
FIGURE 25.3 Choosing a file
Page 6 of 30
Research Training for Social Scientists
SAGE
SAGE Research Methods
2000 SAGE Publications, Ltd. All Rights Reserved.
FIGURE 25.4 The youth cohort study data in the data editor window
Page 7 of 30
Research Training for Social Scientists
SAGE
SAGE Research Methods
2000 SAGE Publications, Ltd. All Rights Reserved.
FIGURE 25.5 Saving a data file
To save a file go to File and then to Save and save in the normal Windows 95 fashion.
When undertaking analysis it is a good practice to have your word processor or a notepad open. To cut and
paste output from SPSS into documents, click on the output see Figure 25.9 go to Edit and the Copy: objects
and then paste as usual in your word processor.
Describing a Single Categorical Variable
The analysis and the comprehension of categorical data, especially data in tables, is a fundamental skill for
social science researchers. Many of the questions which appear on social science surveys collect data that
are categorical and can be measured on a nominal scale. In nominal scales numbers (called codes) are used
to identify an attribute, or category. Numbers provide convenient labels in much the same way as postcodes
label areas. But like postcodes the numbers used on nominal scales have no numerical significance in their
own right. We would never dream of adding two postcodes together! Nominal or categorical scales, are
easy to understand. For example, after having undertaken a survey, if you wanted to distinguish between
Page 8 of 30
Research Training for Social Scientists
SAGE
SAGE Research Methods
2000 SAGE Publications, Ltd. All Rights Reserved.
responses from men and women, for the purposes of analysis, you might want to assign the males to a
category labelled 1 and females to a category labelled 2. The use of numbers in this case does not mean
that 1 is larger than 2. Nor does it imply that 2 is twice as large as 1. Numbers are merely a convenient
way to organize the data. The main issue to grasp is that in nominal scales the numbers used are arbitrary.
They are convenient labels and have no quantitative meaning in their own right. Numbers used in nominal
scales are not ‘measurement’ in the strictest sense of the word. The only mathematical operation that can be
performed on them is to count the number of times they occur (the frequency). The simplest way to describe
a categorical variable is to report the frequencies, or raw scores, which fall into each category of the variable.
A more useful way to describe a categorical variable is to report the percentages that fall into each category
of the variable.
Open the data file ycs.sav (see Figure 25.4).
The Frequencies procedure provides statistics and graphical displays that are useful for describing many
types of variables. For a first look at your data, the Frequencies procedure is a good place to start.
Click on Statistics, then on Summarize, then on Frequencies. You will now be in Frequencies dialogue box.
The variables in the dataset are displayed in the window on the left of the Frequencies dialogue box. It is
possible to scroll up and down the list.
Click on highest RG Social Class (mother or father) and
to move this over to the Variable(s): box.
Click OK. This will produce the frequencies for the highest RG Social Class (mother or father) variable in the
Outputl window (see Figure 25.8).
The first column of the table records the categories of the variable (highest RG Social Class (mother or
father)). These are the familiar categories of the Registrar General’s Social Class Scale.
Page 9 of 30
Research Training for Social Scientists
SAGE
SAGE Research Methods
2000 SAGE Publications, Ltd. All Rights Reserved.
FIGURE 25.6 Selecting the frequencies command
Page 10 of 30
Research Training for Social Scientists
SAGE
SAGE Research Methods
2000 SAGE Publications, Ltd. All Rights Reserved.
FIGURE 25.7 Selecting variables in the frequencies dialogue box
Frequency – This next column of the table reports the frequencies, or number of young people, in each social
class category (e.g. 722 young people have a parent in the professional social class).
Page 11 of 30
Research Training for Social Scientists
SAGE
SAGE Research Methods
2000 SAGE Publications, Ltd. All Rights Reserved.
FIGURE 25.8 Frequency output in the SPSS viewer
Percent – The next column of the table reports the percentage, or proportion, of young people in each social
class category. Using percentages is a very useful method of describing a categorical variable.
Missing values – In this example there are no missing values but in some sets of data there are cases which
do not have a value for a given variable. SPSS will construct a row to represent this category which is labelled
as
Missing in the first column of the table. In the Frequency column the proportion, or percentage, of cases
which have a missing value are reported.
Valid Percent – For some analyses we are only interested in valid cases (i.e. those cases without missing
values). The Valid Percent column of the table reports percentages for valid cases only.
Cumulative Percent – The final column keeps a running total of the percentages. This is sometimes useful as
we can easily deduce that 40.3 per cent of the young people have parents in the intermediate or professional
social classes (i.e. 8.4 per cent + 31.9 per cent).
Page 12 of 30
Research Training for Social Scientists
SAGE
SAGE Research Methods
2000 SAGE Publications, Ltd. All Rights Reserved.
From the Frequencies Dialogue Box it is also possible to construct bar charts, pie charts and histograms by
clicking on the Charts Command.
DESCRIBING CATEGORICAL VARIABLES
Data in tables
The simplest way to represent the relationship between two categorical variables is to produce a table of
frequencies. In the sections below we will be using SPSS to construct a series of tables using the data from
the Youth Cohort Study. The example which we will be looking at is an exploration of the relationship between
gender and participation in post-compulsory education (age 16–17).
The outcome in which we are interested is whether or not a young person has remained in education aged
16–17. This is the Y variable. We are attempting to see if a young person’s gender explains this outcome.
Gender is our explanatory or X variable.
To begin the construction of tables and the analysis of the data you must read in the data from ycs.sav
Click on Statistics, then on Summarize, then on Crosstabs. You will now be in the Crosstabs dialogue box.
(Figure 25.9).
Page 13 of 30
Research Training for Social Scientists
SAGE
SAGE Research Methods
2000 SAGE Publications, Ltd. All Rights Reserved.
FIGURE 25.9 Selecting the crosstabs command
Click on Gender and
to move this into the Columns: box. Click on OK. This will produce a
crosstabulation of Education 16–17 by Gender (Figure 25.10).
The simplest table is a two-by-two table (2 × 2). The raw scores, or frequencies, are hard to interpret
especially when the cell frequencies are large. Calculating percentages is often more useful as they can be
interpreted more easily.
We will now construct a table with column percentages. Click on Window, then click on 1 ycs – SPSS Data
Editor. This will bring you back to the Data Editor window. Click on Statistics, then on Summarize, then on
Crosstabs (Figure 25.11).
Page 14 of 30
Research Training for Social Scientists
SAGE
SAGE Research Methods
2000 SAGE Publications, Ltd. All Rights Reserved.
FIGURE 2S.10 Selecting variables in the crosstabs dialogue box
Page 15 of 30
Research Training for Social Scientists
SAGE
SAGE Research Methods
2000 SAGE Publications, Ltd. All Rights Reserved.
FIGURE 25.11 Crosstabulation of education, 16–17 by gender
It is sometimes difficult to interpret tables with large frequencies in each cell. In these circumstances it is
better to use percentages.
To produce a table with percentages click on Cells at the bottom of the Crosstabs dialogue box. We are now
in the Crosstabs: Cell Display dialogue box. Click on Observed to deselect. Here we select the percentage
required from the Percentages dialogue box at bottom left-hand corner. Click on box beside Column. Click
on Continue and then OK when you return to the Crosstabs dialogue box to run the procedure (Figure
25.10).
We have constructed the table to show column percentages. Conventionally, tables were constructed with the
independent or explanatory variable (X) on the horizontal margin and the dependent variable or outcome (Y)
on the vertical margin. Using this convention and calculating column percentages allows us to interpret the
relationship between X and Y easily (Figure 25.12).
Page 16 of 30
Research Training for Social Scientists
SAGE
SAGE Research Methods
2000 SAGE Publications, Ltd. All Rights Reserved.
FIGURE 25.12 Crosstabulation of education, 16–17 by gender row percentages
The table can now be interpreted more easily than when raw frequencies are used
• 45.5 per cent of the young males in the survey were not in education when surveyed at age 16–17.
• 54.5 per cent of the young males in the survey were in education when surveyed at age 16–17.
• 41.3 per cent of the young women in the survey were not in education when surveyed at age 16–17.
• 58.7 per cent of the young women in the survey were in education when surveyed at age 16–17.
Using this table we can begin to examine the relationship between gender and education in another way
(Figure 25.13).
• 48.8 per cent of the young people in the survey who were not in education when surveyed at age
16–17 were male.
• 51.2 per cent of the young people in the survey who were not in education when surveyed at age
16–17 were female.
• 44.5 per cent of the young people in the survey who were in education when surveyed at age 16–17
were male.
Page 17 of 30
Research Training for Social Scientists
SAGE
SAGE Research Methods
2000 SAGE Publications, Ltd. All Rights Reserved.
• 55.5 per cent of the young people in the survey who were in education when surveyed at age 16–17
were female.
FIGURE 25.13 Crosstabulation of education, 16–17 by gender row percentages
To get the big picture we have calculated total percentages. This is the proportion of cases that fall into a
particular cell. This is sometimes useful when we wish to report the characteristics of a sample (Figure 25.14).
• 21.1 per cent of the young people in the survey were male and not in education at age 16–17.
• 25.3 per cent of the young people in the survey were male and in education at age 16–17.
• 22.1 per cent of the young people in the survey were female and not in education at age 16–17.
• 31.5 per cent of the young people in the survey were female and in education at age 16–17.
Page 18 of 30
Research Training for Social Scientists
SAGE
SAGE Research Methods
2000 SAGE Publications, Ltd. All Rights Reserved.
FIGURE 25.14 Crosstabulation of education, 16–17 by gender total percentages
In this section we have constructed basic tables in SPSS and presented the information in a variety of ways.
The appropriate presentation of the data is very much dependent on what you wish to communicate to the
reader. The data used in tables should be clearly labelled and word processed when presented in research
reports and publications.
Describing Continuous Variables
There are two kinds of measurement scales with which social scientists have to work. We have already come
across categorical data and are about to meet up with continuous data.
Continuous scales are generally ordered, so that it is possible to speak of ‘more’ or ‘less’ of what it is that is
being measured according to the value on the scale.
If there is no true zero (for example, temperature where we have Celsius and Fahrenheit scales with arbitrary
zero points), then we have an interval scale at best. Put simply, when a thermometer using either of these
Page 19 of 30
Research Training for Social Scientists
SAGE
SAGE Research Methods
2000 SAGE Publications, Ltd. All Rights Reserved.
scales reads zero we can’t say that there is no temperature in the room.
If there is a true zero point on a scale then that gives a ratio scale. It is possible to speak meaningfully of a
data point of (say) 20 being twice as high as another data point with a value of 10.
Many continuous measurement scales in social science are interval scales rather than ratio scales. The
difference is captured succinctly by asking the question: ‘Are the data still meaningful if a fixed value (say
50) were added to each score?’ For much data on attitudes or preferences, scales are arbitrary and such an
adjustment would not matter. By contrast data measured in pounds and pence are changed fundamentally by
adding or subtracting a constant. Quantities measured in money terms are usually on ratio scales.
Open the file ycs2.sav using the usual procedure. This is another extract of the Youth Cohort Study of England
and Wales. There are about 2,500 cases in this set of data . The first variable is an identification variable. The
second is the young person’s gender. The next measures the exam grades which the young person obtained
in their final compulsory year of school. The next two variables are related to the young person’s full-time
employment. Pay1 measures their weekly take-home pay age 17–18. Pay2 measures their weekly take-home
pay age 18–19 (Figure 25.15).
FIGURE 25.15 Youth cohort study data file 2
Page 20 of 30
Research Training for Social Scientists
SAGE
SAGE Research Methods
2000 SAGE Publications, Ltd. All Rights Reserved.
Click on Statistics, then click on Summarize, then on Frequencies to display the dialogue box (Figure 25.16).
The variables are listed in the box on the left. Click on the FT job: Weekly take home pay age 17–18 [payl]
variable and then click on the
to select. Make sure that you click on the Display Frequency Tables
box on the left to deselect this option. If you don’t you will get a listing of all the values of the FT job: Weekly
take home pay age 17–18 [payl] variable. In small studies this might be useful, but in large studies like the
YCS, the frequency table is not useful. Click on Statistics at the bottom of the Frequencies dialogue box
(Figure 25.17).
FIGURE 25.16 Selecting the frequency command for contin…
Purchase answer to see full
attachment