Close Window
Last Updated: 25 January 05

Section 1: An Introduction to ChartDog 2.0

ChartDog is a web-based application that allows you to create progress-monitoring graphs. The application follows common conventions for graphing single-subject research data (Hayes, 1981; Kazdin, 1982). With ChartDog, you can:

ChartDog is a useful tool for charting and analyzing outcome data to judge whether a student has made significant improvements in academic or behavioral goals in response to classroom interventions.


Section 2: Getting Started With ChartDog 2.0

When you first open ChartDog 2.0, a 'starter' page wil be displayed (see figure below). The page directs you to select from a drop-down list the number of data points that you want to enter into the application and to click the 'Open ChartDog!' button. Immediately, the page reloads in a format that allows you to enter data to be charted.

ChartDog 2.0: Starter Page

Don't worry if you are not sure how many observations that you will ultimately chart: ChartDog allows you to add additional lines for data entry at any time


Section 3: Customizing ChartDog 2.0 Graphs

ChartDog gives you the ability to customize your data graph to suit your own needs. You can change the following chart settings:

Figure 1: Enter Graph, X-Axis, and Y-Axis Names

Figure 2: Select Line-Plot Options

Figure 3: Show Numeric Values of Data-Points

Figure 4: Select Optional 'Short Names' For Data-Series

Figure 5A: Display Data Table in Report

Figure 5B: Example of Chart Data Table


Section 4: Entering Dates & Data into ChartDog 2.0

For each observation date, you can enter up to two separate data values into ChartDog and select the date on which the observation(s) were collected. ChartDog also allows you to mark multiple phase changes on the data chart and to add custom phase labels.

Figure 7: Enter Dates & Data Values For Each Observation

Figure 8: Example of Labeled Graph

Figure 9A: Create Chart/Erase Chart Data & Settings

Figure 9B: Add Data Lines


Section 5: Saving Data & Chart Settings in ChartDog 2.0

ChartDog 2.0 now allows you to save your chart settings and any data that you have entered into the program. You must complete two steps to permanently save data in ChartDog:

STEP 1: SAVE DATA TO THE WEB PAGE: If no data have yet been saved to the page, you will see a red checked border displayed, along with the message 'No Data Has Been Saved to This Page'(Figure 10A).

To save data, you enter the data you wish to have appear in the chart and customize any chart settings that you wish to save. Then you click the 'Save Data & Chart Settings to This Page' button. The page will reload. A green checked border will be displayed, along with the message 'Your Data Has Been TEMPORARILY Saved to This Web Page' (Figure 10B).

Figure 10A: 'No Data Saved to Page' Message

Figure 10B: 'Data Temporarily Saved to Page' Message

STEP 2: SAVE DATA TO YOUR HARD DRIVE. Please note that your data is not PERMANENTLY saved until you save the page to your hard drive. To save a Web page, click File in the browser's Menu bar. Netscape users, click Save Page As, choose a folder to save the Web page to, and then click Save. IE users, click File and then Save As. Choose a folder and then click Save.

Once you have successfully saved a ChartDog web page with data to your hard drive, you can open that page weeks or months later and all of the data will appear on the page just as you left it. If your computer is connected to the Internet when you open the saved ChartDog page, you can immediately create a chart or add more data and resave the page to the hard drive.


Section 6: ChartDog 2.0 Data Analysis Options

You are most likely to use ChartDog to measure the impact of an intervention, or 'treatment', on a single subject (e.g., a student) over time. If that treatment results in immediate improvements, you may be able to tell just by looking at the pattern of data points on your graph that your intervention has proven highly effective. However, on many time-series graphs, it is typical for the data to vary considerably from day to day. When faced with so much 'scatter' or variability in the data points on your chart, you may find it difficult to pick out any strong underlying pattern or trend through visual inspection alone. ChartDog gives you several computational and statistical tools that you can use to help you to analyze and draw conclusions from your charted data.
By selecting data-analysis tools from the ChartDog drop-list (Figure 11), you can:

Figure 12: Example of a ChartDog Time-Series Chart Report

You can choose to use any or all of these tools to analyze data from a particular graph. The results of each analysis appear on the chart report, below the time-series graph (see sample chart above in Figure 12). Each data-analysis tool has its own strengths and limitations--as described below:


'Compute Trend (Regression) Lines for Data Series...' When drawn on a graph, trend lines summarize the direction and rate of change in a data series (Franklin, Gorman, Beasley, & Alison, 1996). ChartDog calculates trend by using ordinary least-squares (OLS) regression (Moore & McCabe, 1989), a statistical formula that is widely used in time-series graphs (Franklin, Gorman, Beasley, & Alison, 1996), including those used for curriculum-based measurement monitoring (Shinn, Good, & Stein, 1989). As shown in Figure 12, the computation of OLS regression results in slope and intercept values. Using these values, ChartDog draws a dotted-line through the data series that shows the best 'linear fit' of all values (see example in Figure 14A).


Figure 13: Formula for Computing Least-Squares Regression

Figure 14: Chart Examples of Regression Analysis (A) and Mean Value (B)

You can view the trend line to get a visual estimate of the direction and rate of improvement for the individual whose data are being graphed. Slope and intercept values also appear in the 'Graph Notes' window.

Cautions and Limitations in Interpreting OLS Regression:


'Compute Means for Data Series...' For each phase in your graph, ChartDog can compute the mean value of data points present within that phase. When you select this option, you see that mean values are represented on your graph as horizontal lines within each phase (see Figure 14B for an example). Means for each data series are drawn in the same color that you select for that data series (e.g., green, red, blue). Mean values also appear in the 'Graph Notes' window.

Cautions and Limitations in Interpreting Phase Means:


'Compute Percentage of Non-Overlapping Data Points (PNDs) for Data Series...' One method for quantifying the impact of a treatment or intervention phase in a data series is to calculate the percentage of non-overlapping data points (PNDs) in the treatment phase when compared to the baseline phase (Faith, Allison, & Gorman, 1996); Scruggs, Mastropieri, & Casto, 1987). PNDs are computed in four steps:

  1. Decide whether the target behavior that you are measuring is intended to increase or to decrease during the treatment phase.
  2. Count up the number of data points in the treatment phase that do not overlap the data points in the baseline phase.

  3. Divide the number of non-overlapping data points in the treatment series by the TOTAL number of data points in the treatment series.
  4. Multiply the figure from step 3 by 100.

A brief example will illustrate the calculation of PNDs. Figure 15 shows a series of 4 data points (treatment phase) being compared to a series of 3 data points (baseline phase). Higher data points represent the desired direction for improvement in this data series (Step 1). There are 4 data points in the treatment phase that do not overlap baseline data (Step 2).

Figure 15: Example of PND Calculation

We find that there are also a total of four data points in the entire treatment series. When we divide 4 non-overlapping data points by the total of 4 values in the series, we get a quotient of 1.0 (Step 3). We then multiply this quotient by 100 to find that the PND for the treatment phase is 100%. ChartDog displays PND values in the 'Graph Notes' window.

Cautions and Limitations in Interpreting PNDs:


'Compute Effect Sizes for Data Series...' Advanced users of ChartDog may wish to calculate a standardized 'effect size' for treatment data collected on an individual subject. By calculating effect sizes for each subject in a multi-participant study, a researcher can then report the mean of those values as the 'average' effect size, or outcome, of the study. Or ChartDog users may want to convert the findings of a single-subject study to an effect-size so that they can compare their results to published studies of a similar nature.

While a number of formulas exist for calculating effect sizes in single-subject designs, ChartDog employs a commonly used 'standardized difference approach' (Faith, Allison, & Gorman, 1996; Shernoff, Kratochwill, & Stoiber, 2002). Using this formula (Figure 16):

Figure 16: Formula for Computing Effect Size Between Treatment and Baseline Phases (Faith, Allison, & Gorman, 1996)

When you prompt ChartDog to calculate effect sizes, the program begins with the second phase and cycles through all phases appearing in the chart. Treating the current phase as the 'treatment' phase and the phase preceding the current phase as the 'baseline' phase, ChartDog successively calculates an effect size for each treatment phase. Effect-size values appear in decimal format and are displayed in the 'Graph Notes' window.

Cautions and Limitations in Interpreting Effect Sizes (Faith, Allison, & Gorman, 1996):


Section 7: About ChartDog 2.0

ChartDog is a web-based program that constructs customized progress-monitoring time-series graphs from data-values entered by the user. It is programmed in PHP, an Internet [computer] scripting language. ChartDog is built upon the foundation of JPGraph, a freeware suite of PHP code that creates dynamic charts for the Internet. JPGraph was created by Johan Persson, a Swedish programmer who has demonstrated great generosity in making his software available at no cost to web developers. Learn more about JPGraph at:

The dog icons used throughout the ChartDog web form and manual came from Nory Island, a Japanese web site with neat graphics.

I (Jim Wright) would also like to express my appreciation to some folks who helped me to create or improve ChartDog, including:


Dunn, E.K. & Eckert, T.L. (2002). Curriculum-based measurement in reading: A comparison of similar versus challenging material. School Psychology Quarterly, 17, 24-46.
Faith,M.S., Allison, D.B., & Gorman, B.S.(1996). Meta-analysis of single-case research. In R.D.Franklin, D.B.Allison, & B.S.Gorman (Eds.), Design and analysis of single-case research. (pp.245-277). Mahwah, NJ: Lawrence Erlbaum Associates.
Franklin,R.D., Gorman, B.S., Beasley, T.M., & Allison, D.B. (1996). Graphical display and visual analysis. In R.D.Franklin, D.B.Allison, & B.S.Gorman (Eds.), Design and analysis of single-case research. (pp.119-158). Mahwah, NJ: Lawrence Erlbaum Associates
Hayes, S.C. (1981). Single case experimental design and empirical clinical practice. Journal of Consulting and Clinical Psychology, 49, 193-211.
Kazdin, A.E. (1982). Single-case research designs: Methods for clinical and applied settings. New York: Oxford Press.
Moore, D.S., & McCabe, G.P. (1989). Introduction to the practice of statistics. New York: W.H. Freeman and Company.
Scruggs, T.E., Mastropieri, M.A., & Casto, G. (1987). The quantitative synthesis of single-subject research: Methodology and validation. Remedial and Special Education, 8(2), 24-33.
Shernoff, E.S., Kratochwill, T.R., & Stoiber, K.C. (2002). Evidence-based interventions in school psychology: An illustration of task force coding criteria using single-participant research design. School Psychology Quarterly, 17, 390-422.
Shinn, M.R., Good, R.H., & Stein, S. (1989). Summarizing trends in student achievement: A comparison of methods. School Psychology Review, 18, 356-370.

Close Window