Specifically, computation items can be used to:<\/p>\n\n\n\n
- \n
- Compute a sum score or average score of a questionnaire.<\/li>\n\n\n\n
- Use advanced conditional logic.<\/li>\n\n\n\n
- Compute new constructs that result from the client’s previous answers.<\/li>\n\n\n\n
- Randomize questions.<\/li>\n<\/ul>\n\n\n\n
![\"\"](\"\/wp-content\/uploads\/2021\/10\/compu8.png\")
dashboard-5 or higher should be used<\/figcaption><\/figure><\/div>\n\n\n Pseudo R<\/h2>\n\n\n\n
The scripting language used to write computations is based on R. Therefore it’s called Pseudo R. R was chosen as this is probably the language most psychologists or even statisticians are familiar with. The goal is to implement as many functionalities from R to make real-time computations in m-Path as useful as possible.<\/p>\n\n\n\n
Let us know (contact@m-path.io<\/a>) if you are still missing some functionalities. We are happy to implement them. <\/p>\n\n\n\n
Example questionnaire ‘computation examples’<\/h2>\n\n\n\n
Most computations in this article can be tested by using the ‘computation examples R<\/strong>‘ questionnaire from Freud which can be found in the m-Path questionnaire library. This questionnaire starts with two questions:<\/p>\n\n\n\n
- \n
- How happy are you now? (label ‘happy’)<\/li>\n\n\n\n
- How sad are you now? (label ‘sad’)<\/li>\n<\/ul>\n\n\n\n
In the last question, all answers from all computation questions are shown using piped text<\/a>.<\/p>\n\n\n\n
How to create a computation item?<\/h2>\n\n\n\n
- \n
- Choose the Computation question type.<\/li>\n\n\n\n
- Choose the label of the computation, for example ‘happyPlus50’.<\/li>\n\n\n\n
- Make sure to set ‘code script<\/strong>‘ in pseudo R<\/em>. If you are not able to set the script to pseudo R, you are still working with an older dashboard. The old computation item is explained here<\/a>.<\/li>\n\n\n\n
- Type the Mathematical expression<\/em>, for example happy+50<\/em>. The value of ‘happyPlus50’ will then be set to 50 plus the answer of the question with label ‘happy’.<\/li>\n\n\n\n
- [optional] Fill in the minimum<\/em> and maximum value for visualization<\/em>. If this computation question needs to be visualized in a plot, m-Path needs to know its minimum and maximum values. As the minimum and maximum value of the ‘How happy are you now?’ question are 0 and 100, the minimum and maximum value of happyPlus50 are 50 and 150.<\/li>\n<\/ol>\n\n\n\n
![\"\"](\"\/wp-content\/uploads\/2021\/10\/compu1-1.png\")
A simple computation question in the ‘computation examples’ questionnaire.<\/figcaption><\/figure>\n\n\n\n Possible mathematical expressions<\/h2>\n\n\n\n
Normal mathematical expressions<\/h3>\n\n\n\n
A normal mathematical expression is an expression such as 5+3*2 +4\/2<\/em>. This will give 13 as a result (see question with label normalMathExpression_1 in example). An absolute value is also possible. For example (-5)^2<\/em> will result in 25 (see questions with label ‘normalMathExpression_2’).<\/p>\n\n\n\n
Use answers from previous questions<\/h3>\n\n\n\n
Answers from previous questions (from the same questionnaire) can also be used in mathematical expressions. To access these answers, the label of the previous questions can be used in the expression. For example happy-sad<\/em> can be used to subtract ‘sad’ from ‘happy’ to get some general affect (see label question with label ‘happyMinusSad’ in example).<\/em><\/p>\n\n\n\n
Labels from previous computation questions can also be used (E.g. happy+happyPlus50<\/em> in question with label ‘happyPlus50Re-use’).<\/p>\n\n\n\n
![\"\"](\"\/wp-content\/uploads\/2021\/10\/compu3.png\")
How to use answers from previous questions<\/em>.<\/figcaption><\/figure>\n\n\n\n Make sure all questions needed for the expression are answered, otherwise the answers can not be found and the app will crash.
E.g. If the ‘happy’ question is only answered in some conditions, [happy] should not be used in computation questions outside this condition.<\/em><\/p>\n\n\n\nScripts or multiline mathematical expressions<\/h3>\n\n\n\n
Like in R, it is also possible to create expressions that span over multiple lines (so-called scripts). In each line a new variable can be declared. Declaring a variable can be done using the = <\/em>or the <-<\/em> sign. In the following script (found in question with label ‘scriptExample’) three new variables will be set:<\/p>\n\n\n\n
scriptComputation1<-100\nscriptComputation2=happy*4\nscriptComputation3=scriptComputation2-100<\/code><\/pre>\n\n\n\nThis script will result in 3 answers (with labels ‘scriptComputation1’, ‘scriptComputation2’ and ‘scriptComputation3’). All of them can be used in later questions. <\/p>\n\n\n\n
Note that ‘scriptExample’ is the label of the question, but the variable itself is not set. There will only be an answer with label ‘scriptExample’ if<\/p>\n\n\n\n
- \n
- There is an ans expression. This is an expression where NO new variable is explicitly set, like 5*happy<\/em> (but not newVariable = 5*happy<\/em>). The answer of the overall label of the question will be set to the last ans expression.<\/li>\n\n\n\n
- A variable is set with label ‘scriptExample’. For example scriptExample=70<\/em>.<\/li>\n<\/ul>\n\n\n\n
![\"\"](\"\/wp-content\/uploads\/2021\/10\/compu4.png\")
A script of multiple expressions<\/figcaption><\/figure>\n\n\n\n Comments in script<\/h3>\n\n\n\n
It is also possible to add comments in the script. Everything after #<\/em> will be ignored.<\/p>\n\n\n\n
# this statement will be ignored in the script <\/code><\/pre>\n\n\n\nVectors or arrays<\/h3>\n\n\n\n
It is possible to create an array using the c()<\/em> command. The following script will create a new array and subsequently add a number to it:<\/p>\n\n\n\n
arrayExample=c(1,2,3,4)\narrayExample=c(arrayExample,5) <\/code><\/pre>\n\n\n\nYou can subsequently access an array using []<\/p>\n\n\n\n
secondElement=arrayExample[2]\nsecAndThirdElement=arrayExample[c(2,3)]\n#or with an id vector\nid=c(2,3)\nsecAndThirdElement=arrayExample[id]<\/code><\/pre>\n\n\n\n<\/p>\n\n\n\n
Strings<\/h3>\n\n\n\n
You can define strings using single or double quotations marks.<\/p>\n\n\n\n
text1 = 'Text1'\ntext2 = \"Text2\"<\/code><\/pre>\n\n\n\nUsing functions<\/h3>\n\n\n\n
Many functions from R can already be used in m-Path. Only the base functions can however be used, additional settings can not be included. For example mean(c(1,2,3))<\/em> <\/em>is possible but mean(c(1,2,3),trim=0.1)<\/em> is not. <\/p>\n\n\n\n
Additionally, also some functions that are not in R are included. For example, the reverse <\/em>function reverses the score, based on the scale of the variable. This can be very useful when creating a construct that includes reverse scores. <\/p>\n\n\n\n
All functions are shown in the example questionnaire in the question with label ‘allFunctions’.<\/p>\n\n\n\n
- \n
mean(c(happy,sad))<\/code>: the mean of ‘happy’ and ‘sad’.<\/li>\n\n\n\nmedian(c(happy,sad))<\/code>: the median of ‘happy’ and ‘sad’.<\/li>\n\n\n\nnanmean(c(happy,sad,other))<\/code>: the mean of ‘happy’, ‘sad’ and ‘other’. Exclude elements that do not exist or are na (which is the case for ‘other’ in the example). At least a single element has to exist for this function to work. This function is equivalent tomean(c(happy,sad,other), na.rm=TRUE)<\/code> in R.<\/li>\n\n\n\nmax(c(happy,sad))<\/code>: the maximum of ‘happy’ and ‘sad’.<\/li>\n\n\n\nmin(c(happy,sad))<\/code>: the the minimum of ‘happy’ and ‘sad’.<\/li>\n\n\n\nsum(c(happy,sad))<\/code>:<\/em> for the sum of ‘happy’ and ‘sad’.<\/li>\n\n\n\nsqrt(happy)<\/code>: the square root of ‘happy’<\/li>\n\n\n\nlog(happy)<\/code>: the log of ‘happy’<\/li>\n\n\n\nabs(happy)<\/code>: the absolute value of ‘happy’<\/li>\n\n\n\nceil(happy)<\/code>: the value of ‘happy’ ceiled<\/li>\n\n\n\nfloor(happy)<\/code>: the value of ‘happy’ floor<\/li>\n\n\n\nround(happy)<\/code>: the value of ‘happy’ rounded<\/li>\n\n\n\nquantile(c(happy,sad),0.5)<\/code>: the 50% percentile of the array.<\/li>\n\n\n\nlength(c(happy,sad))<\/code>: the length of the array<\/li>\n\n\n\nrnorm(3)<\/code>: creates an array of 3 random variables (normally distributed)<\/li>\n\n\n\nrunif(3)<\/code>: creates an array of 3 random variables (uniformly distributed)<\/li>\n\n\n\nis.na(happy)<\/code>: returns 1 if happy is not available and 0 if it exists<\/li>\n\n\n\nreverse(happy)<\/code>: reverses the score of happy. In this case, as happy is on a scale from 0 to 100 it will result in 100-happy. If it would have been a seven point likert scale, it would have resulted in 7-happy+1 (such that the reverse score goes from 7 to 1).<\/li>\n\n\n\nas.character(happy)<\/code>: converts a number to a string.<\/li>\n\n\n\nnchar('test')<\/code>: counts the characters in a string. This example will return 4.<\/li>\n\n\n\npaste('test','1')<\/code>: this will paste two strings together. This example will result in ‘test1’.<\/li>\n\n\n\nsubstr('test5',2,3)<\/code>: this will generate a substring. This example will result in ‘es’.<\/li>\n\n\n\ntolower('Test')<\/code>: transforms a string to lower case. In this case the result would be ‘test’.<\/li>\n\n\n\ngrepl('pattern','string that contains pattern')<\/code>: checks if a string contains a pattern. In this case the output would be 1 (true).<\/li>\n\n\n\nas.Date(time)<\/code>: this will format a time (seconds since epoch) into a date format. It outputs a string in format ‘year-month-day’. It will also transform UTC time to current time. Use as.Date(Sys.time)<\/em> for the current date.<\/li>\n\n\n\nas.weekday(time)<\/code>: this will return what day of the week this timestamp indicates in an integer value (with Monday being 1 and Sunday being 7).<\/li>\n\n\n\nas.day(time)<\/code>: this will return what day of the month this timestamp indicates in an integer value.<\/li>\n\n\n\nas.hours(time)<\/code>: this will format a time (seconds since epoch) into the hours of a day. It will also transform UTC time to current time. Use as.hours(Sys.time)<\/em> for the current hours of the day. Also see example 8.<\/li>\n\n\n\nas.minutes(time)<\/code>: this will format a time (seconds since epoch) into the minutes of a day. It will also transform UTC time to current time. Use as.minutes(Sys.time)<\/em> for the current minutes of the day. Also see example 8.<\/li>\n\n\n\nas.seconds(time)<\/code>: this will format a time (seconds since epoch) into the seconds of a day. It will also transform UTC time to current time. Use as.seconds(Sys.time)<\/em> for the current seconds of the day. Also see example 8.<\/li>\n\n\n\nas.epochTime(<\/code>“time<\/code>“)<\/code>: This will calculate the corresponding time (seconds since epoch) into an integer value. Example: as.epochTime(“9:00”) will return the seconds since epoch value for the last time<\/strong> it was 9:00.<\/li>\n\n\n\nis.skipped(happy)<\/code>: returns 1 (true) if question with label happy <\/em>was skipped. This is useful for questions that can be skipped<\/a>. See ’empty answer test’ questionnaire for examples.<\/li>\n\n\n\nis.timedout(happy)<\/code>: returns 1 (true) if question with label happy <\/em>was timed out. This is useful for questions with a deadline<\/a>. See ’empty answer test’ questionnaire for examples.<\/li>\n\n\n\nis.empty(happy)<\/code>: returns 1 (true) if question with label happy <\/em>was empty. If a question was not yet filled in and did not have a valid starting value when it was skipped or timed out, the answer is empty. See ’empty answer test’ questionnaire for examples.<\/li>\n\n\n\nmatch(a, b)<\/code>: returns a vector of the positions of (first) matches of its first argument in its second. For example match(3,c(1,4,5,3,6))<\/em> will return 4.<\/li>\n\n\n\nas.responsetime(happy)<\/code>: gives the response times of the question with label happy<\/em>. Using this feature, it is possible to in real time adapt a questionnaire to a participant that is answering to quickly. <\/li>\n\n\n\nas.question(happy<\/code>): gives you the question related to the question with label happy. For example ‘how happy are you now?’.<\/li>\n\n\n\nsample.int(5,1)<\/code>: samples from an integers. In this case, one integer is generated from the numbers 1 to 5. In case of sample.int(5,2)<\/em>, two integers are generated (without replacement).<\/li>\n\n\n\nshewhart.UCL(a)<\/code>: if a<\/em> is a list or vector of numbers, the shewhart.UCL(a)<\/em> gives you the upper bound for Shewhart procedures of process control.<\/a> <\/li>\n\n\n\nshewhart.LCL(a)<\/code>: if a<\/em> is a list or vector of numbers, the shewhart.LCL(a)<\/em> gives you the lowerbound for Shewhart procedures of process control.<\/a><\/li>\n\n\n\nseq(1,5)<\/code>: gives you a sequence of integers 1 to 5.<\/li>\n\n\n\nvar(a)<\/code>: if a is a list or vector of numbers, var <\/em>gives you the variance of these numbers<\/li>\n\n\n\nsd(a)<\/code>: if a is a list or vector of numbers, sd <\/em>gives you the standard deviation of these numbers<\/li>\n\n\n\ndistm(a,b)<\/code>: if a<\/em> is a vector with a latitude and a longitude and b <\/em>is a vector with a latidude and a longitude, distm <\/em>will give you the distance between both points using the Haversine formula in meters. For example distm(c(51.04, 4.488),c(52.14, 3.523)).<\/em> This can be useful if a location sensor is used.<\/li>\n\n\n\nwhich(a)<\/code>: returns the position or the index of the value for which a is equal to 1. For example which(c(1,0,1,0)) <\/em>would return c(1,3)<\/em>. As such which(happyList>50)<\/em> returns a list of indexes where the numbers in happyList are larger than 50.<\/li>\n\n\n\nsort(a)<\/code>: sorts this vector from low to high.<\/li>\n\n\n\nunique(a)<\/code>: Returns a vector with all the unique values from a.<\/li>\n\n\n\ngetResponseTime(question_label)<\/code>: Returns the total time (in milliseconds) the participant spent answering a specific question.<\/li>\n\n\n\ngetLastChange(question_label)<\/code>: Returns the time elapsed (in milliseconds) from when the participant started the question until they made their final answer change..<\/li>\n\n\n\ngetMultipleChoiceAnswer(question_label)<\/code>: return a vector containing the string answers for a multiple choice question with label question_label.<\/li>\n<\/ul>\n\n\n\nIt is impossible to use labels with the name of a function in computation items. Choose a label that is not a function. For example mean=3<\/em> will not work.<\/p>\n\n\n\n
Combination of the above<\/h4>\n\n\n\n
Use the ()<\/em> brackets to work with nested expressions. For example statement mean(c(max(c(happy,70)),sad))*2<\/em> can be found in the example questionnaire under the question with label ‘bracketExample’.<\/p>\n\n\n\n
Logical expressions<\/h3>\n\n\n\n
The following signs can be used to create a logical expression or an equation: >,<,==,<=<\/em> and >=<\/em>. If the expression is true, the result will be 1, if not, the result will be 0. Additionally, logical operators like &&<\/em> and ||<\/em> can be used.<\/p>\n\n\n\n
For example happy>sad && sad > 50 <\/em> (see question with label ‘logicalStatement’ in example questionnaire) will result in 1 only if the answer of ‘happy’ is bigger than the answer of ‘sad’ and ‘sad’ is higher than 50.<\/p>\n\n\n\n
If else statements<\/h3>\n\n\n\n
It is also possible to use if else statements in the script. The following will set ‘ifElseVariable’ to 1 if happy is bigger than sad. Otherwise it will be set to 0 (see question with label ‘ifElseVariable’).<\/p>\n\n\n\n
if(happy>sad){\n ifElseVariable=1\n}else{\n ifElseVariable=0\n}<\/code><\/pre>\n\n\n\nConstants or predefined variables<\/h3>\n\n\n\n
It is possible to call predefined variables:<\/p>\n\n\n\n
- \n
Sys.time<\/code>: the current time in seconds. This is in UTC time. If the current hour in local time is needed, one can use as.hours(Sys.time)<\/em> (see functions above).<\/li>\n\n\n\nsystem_nToday<\/code>: The number of notifications answered on that day by the user, including the current questionnaire. If this is equal to one, it means that the user answering the first notification of the day. If the same notification is answered twice, this number will not increase. This includes all scheduled notifications and follow-up notifications (but not manually sent notifications).<\/li>\n\n\n\nsystem_nTodayL<\/code>: The number of questionnaires filled in on that day by the user, including the current questionnaire. If the same notification is answered twice, this number will increase.<\/li>\n\n\n\nsystem_nThisWeek<\/code>: The number of notifications answered in the previous week. If the same notification is answered twice, this number will not increase. This includes all scheduled notifications and follow-up notifications (but not manually sent notifications).<\/li>\n\n\n\nsystem_progToday<\/code>: the percentage of notifications answered by the participant on that day. Equals to system_nToday divided by the total number of notifications that are expected to be sent on that day.<\/li>\n\n\n\nsystem_maxstreak<\/code>: The amount of consecutive days the participant has filled in their questionnaires.<\/li>\n\n\n\nSys.alias<\/code>: the current nickname of the participant.<\/li>\n\n\n\nSys.connectionId<\/code>: the connectionId of the participant in the context of this interaction.<\/li>\n<\/ul>\n\n\n\nConditional on computation question<\/h2>\n\n\n\n
Conditions can be defined for computation questions. The interactions or questions in the yes<\/em> condition will only be shown if the result of the computation question was equal to 1. More specifically, it will check the answer from the label of the computation question. In the following example, this label is conditionalComputation<\/em><\/em>. <\/p>\n\n\n\n
If conditionalComputation<\/em> is equal to 1, the yes condition will be triggerd. If conditionalComputation<\/em> is 0, the no condition will be triggerd.<\/p>\n\n\n\n
![\"\"](\"\/wp-content\/uploads\/2024\/04\/Capture.jpg\")
- A variable is set with label ‘scriptExample’. For example scriptExample=70<\/em>.<\/li>\n<\/ul>\n\n\n\n
- Type the Mathematical expression<\/em>, for example happy+50<\/em>. The value of ‘happyPlus50’ will then be set to 50 plus the answer of the question with label ‘happy’.<\/li>\n\n\n\n