Using Excel and SPS in a Major Research Paper - Assignment Example

Using Excel and SPS in a Major Research Paper For a successful Major Research Paper, reliable and accurately analyzed statistics are essential. Itis thus necessary that you make the best use of the data you are able to collect, and ensure that the spreadsheet program you use for analysis gives you all the tools you need for accurate, effective analysis. Two packages are examined in these notes. Consider the nature of your own data, and the results you will need from your analysis – you will have to support and explain the claims you make in your research paper. Remember that you will have to communicate your findings clearly to a reader, so the analysis of your data and subsequent presentation of your findings must be clear and accurate. The data analysis package you use must be able to help you fulfill these requirements. Microsoft Excel allows the capture of a large amount of data. This program lets you enter data into 1 million rows, in 16 000 columns. If you were to use all the “space” available to you in a Microsoft Excel spreadsheet, you may need extra tools to help you manage and analyze such huge volumes of information. One tool you might consider using is IBM SPSS Advantage for Excel. This package extends the capabilities of Excel 2007 by including specialized tools in the package. All your data management, Access, and analysis needs may be met by this program. Importantly, even if you do not have extensive personal knowledge of Statistics, the SPSS statistical techniques will help you understand and work with your data. For example, you can examine recency and frequency within this program, guided entirely by the Wizards it features. Hence, while you can manage and explore your data fairly easily with the Wizards, you can also find values in your spreadsheets, and analyze your data, no matter the size of your spreadsheet. In my opinion, using the SPSS functionalities helps to perform a number of tasks more efficiently. The following list of comments will explain my meaning more clearly: In Excel, descriptives, statistical requirements, and frequency data can be managed with some of the program’s standard functions. SPSS enables this analysis to take place much more quickly and simply. The SPSS add-on has a far larger number of available charts and graphs. This is a valuable feature, but I must add that Excel charts and graphs may be easier to manage for some people. Excel Pivot Tables are easy to manipulate. The SPSS tables, though, give you access to far more functions. They are highly flexible and easy to use. It is sometimes difficult to find full sets of statistical tests in Excel, and working with these tests can be challenging. An advantage in SPSS is that statistical tests mostly run at the same time as every other function you use. Similarly the “split” function in SPSS enables you to create reports and graphics very easily. This split capability lets you create tables and charts without any extra capturing of data or additional analysis. You are also able to create syntax (SPSS’s Macro-language) which enables the re-use of tables and codes, so that it is much easier to work with subsets of your data. A very valuable feature in SPSS is the ability to use labels in reports. Instead of codes: for example Male=1; Female=2, you can label your data. This means that your information is not filled with number codes (which you can change to words in Excel). Instead, your values remain intact, but have labels attached to them to describe them. There is a missing value analysis module within SPSS which can be purchased (for approximately $800). By using this function accurate results can be achieved even when some values are missing. You can establish how questions that were skipped by respondents will impact on your final results. You can even get the program to estimate what missing data was likely to have been. This is a definite improvement on what can be done in Excel, but the expense of the SPSS add-on must be taken into consideration. Both Excel and SPSS help you to recognize any errors in data entry or any unusual data points. Neither of these programs has a clear advantage over the other in this aspect. Excel has a smooth, efficient interface and process for import and export of external documents. SPSS allows import of text files, ODBC data bases, and even Excel spreadsheets, Access tables and its own SPSS data bases. The SPSS interface, however, seems to be a little less user-friendly than that of Excel. Both programs easily allow you to manage unlimited numbers of rows. Statistical analysis is generally easier and faster using SPSS. There are, though some areas of analysis for which I prefer using Excel. “Help” files and Tutorials available in SPSS do assist in the user’s understanding of statistical results. These Help and Tutorial functions advise the user on interpretations and with the jargon used in the program. This kind of assistance is not available at all in Excel. As I mentioned, though, there are a number of functions in Excel which I find very useful. Consequently I still use the Excel program for a number of things: Learning to use Excel is reasonably simple. Learning SPSS, however, was at first a very unpleasant experience. Many of the vast number of options are difficult to learn about, and you have to spend a fair amount of time in the program before all these functions become familiar, Once you are completely comfortable within the program, though, SPSS is indispensible, The expense of the SPSS package could certainly put some people off. If you already have Excel, you may not want to spend the money – over $1 700 for the full package. Charts are easier to manage and manipulate in Excel. SPSS has many excellent features in this regard, but in my opinion they are less dynamic than what can be found in Excel. It is often necessary to return to charts before they can be presented or published. I need to rearrange data, to tweak charts, to redefine and alter bars frequently. In Excel, this is easy. You simply edit the underlying spreadsheet, and changes are drawn to the chart and to a PowerPoint presentation automatically. When using SPSS, the chart has to be recreated and subsequently copied into a presentation. For me, Excel has very flexible functions and many more functions than SPSS. Such flexibility means that I do prefer to use Excel when the data requires it. Now, following, are more specific guidelines to using each of these programs. You should decide which software would most suit your needs and skills levels. Remember that it is essential for you to present the findings of your Major Research paper in a clear and accessible way – your reader should not have to struggle to understand your statistics. Using Excel Once you have collected all your data, you are faced with a large set of numbers that need to be analyzed. Your first priority should be to learn more about the distribution of the data, and to examine such factors as minimum and maximum values, or outliers. These three bits of information, particularly, will assist in the evaluation of your data, to check whether they will meet requirements for any further analysis, such as t-tests, or regression. Examine the following graphic: DIAGRAM The representation of an Excel screen shows how you would use Excel to examine the variable “AGE”. Your process is as follows: 1. Select Tools/Data Analysis/Descriptive Statistics. (Be sure that you have the “Data Analysis” option on your Tools menu in Excel. You may have to install the option using Tools/Add-ins …) 2. Select the input range – for our example, it is the AGE variable, as indicated ($B$2:$B$51). 3. Check the boxes: “Summary Statistics” and “Compliance Level” for mean. If you receive a 95% measure, your data are acceptable. The output you will receive is shown in the diagram below: DIAGRAM Important aspects of this display must be noted in your early analysis: 1. Outliers: Find the minimum and maximum values and evaluate whether they are within your expected range for the data. Should a particular value be unexpectedly small or large, examine the original data and check your coding. 2. Symmetry: find the “Skewness” and “Kurtosis” measurements on the display. These will help you to determine whether your data are normal. Skewness will tell you whether the data are non-symmetric. The closer this value is to “0” the more symmetrical the data are. You can measure the standard error of skewness by calculating: =SQRT(6/N) or SQRT(6/50) – which gives you the value: 0.346. If the skewness measure is more than twice this number, the distribution must be considered non-symmetric. 3. Kurtosis: this measures the peakedness of the data. Normally distributed data will produce a measure of kurtosis = 0. A formula for the standard error of kurtosis can be calculated at –SQRT(24/N) thus equal to 0.692. Double this number is 1.39. Hence the data represented here can be considered normal by this measure – the kurtosis value falls within the standard error (-0.26). Further, you should try to visualize your data – this will help you understand it more fully. Select Tools/Data Analysis/ Histogram and select the AGE variable ($B$2:$B$22). Click on the option “Chart Output”. The histogram below will be your result: DIAGRAM Using SPSS Start SPSS for Windows by clicking on the Start menu and select Programs. Choose SPSS 15.0 for Windows. Double-click on the SPSS 17.0 icon on your Desktop. The SPSS window will open but note that SPSS opens consecutive windows as operations are performed. The most used of these will be the Data Editor and the Output Viewer windows. Note also that each new window will have a title bar and a menu bar. Once you have opened the SPSS window, two options are possible: Creating a new data set OR Opening an existing data set See the graphic below to familiarize yourself with the Data Editor window. DIAGRAM Data Editor as you can see is a spreadsheet. You are able to define your variables here and enter your data values. The title bar will indicate the name of your open, saved data file. If you have not yet saved the file, “Untitled” will display. You can work on any of the displayed windows once you have selected it by clicking anywhere on it or by selecting it from the Window menu to make it active. Keystrokes and commands will be possible only in the active window – you will also see that the title bar changes color to indicate that a particular window is active. An active cell in a spreadsheet has a darker border, and is the only one in which typing can be done. To make a cell active, click on it or use tab, return key, or cursor keys to move the dark border to a selected cell. The spreadsheet within the Data Editor has rows which correspond to a case or subject. The columns represent variables. Below is a representation of the Variable Viewer screen: DIAGRAM This screen allows you to define variables for your spreadsheet. These can be defined in a number of ways: 1. Description of Variables Use the Variable Viewer dialog box Access is gained by left-clicking on the Variable Viewer All your variables will be listed and defined here, including labels, type, missing values and so on. The Variable Description area displays all information relating to a particular variable. A default description is available – which is always a numeric variable with a column width of 8, right-justified and without decimal places, and with no labels or missing values. For most of your data, this default description should be enough for most variables. Should you need to change the variable description, use Settings. In my experience, you most frequently may want to change only the labels for your variables. 2. Naming of Variables The “Variable Name” field displays the name of a specific variable. Names can be edited by placing the cursor directly on the field it is in. It is advisable to use names for your variables that are most relevant to your research. Variable names must begin with a letter of the alphabet and may not exceed 60 characters. (Earlier versions of the program only allow for 8 characters.) No spaces are permissible within the variable name and make sure that your names are each unique – i.e. not repeated at all. My advice is that you make your names easy to remember – identity, gender of age are easy to remember and identify. 3. Labels for Variables Two types of labels can be accommodated: Variable Labels – “long names” for the variable. While particular variable names are limited to 60 characters, Variable Labels can be up to 256 characters and can include spaces. Importantly, entering a variable label will allow this label to be put on your charts and reports. This makes them easier to understand. Value Labels – translate numerical data with a key, so that they become understandable and recognizable words. If numbers, for example, “1” represents “Male” and “2” “Female”, the value labels will assist you to recognize the data. Read More