17_Supplemtrary Coding Notes_D1

 Dimension 1: Joint Attention  

Joint attention is coded in ELAN, in several steps – from identification to annotation and labelling. This supplementary coding blog outlines these coding steps and should be used alongside the coding manual, as well as coded video examples (and only after you feel competent with ELAN). 

Note, the most recent training video 007, 14m (and subsequent uncoded files) are the most updated coding templates and reflect the information documented in this supplementary blog.

The coding steps outlined in this supplementary coding blogpost include: 

1. Coding file set-up: annotation mode

2. Identifying joint attention cues

3. Coding joint attention cues

  3.1. Evaluating interaction windows  

  3.2. Labelling interaction windows  

1. Coding file set-up: annotation mode 

First, the coding file should be opened in the correct coding mode – i.e., annotation mode. This is the mode whereby the majority of coding will take place (see Figure 1, below). 

Figure 1: Image depicting the annotation mode

2. Identifying initiations

Within the annotation mode, start by selecting and playing short intervals at a time (e.g., 10-seconds). This is important since many gestures can occur within a few seconds of the interaction; thus, focusing on short intervals increases the detection rate. 

Starting from the beginning of the video, select short 10-second intervals of the interaction by dragging the media crosshair from left to right and then play that interval. Next, visually identify any possible initiation requests which may be present. 

Figure 2: Image depicting the crosshair area and selection media controls.

3. Marking initiation requests

Once an initiation request has been visually identified in the interaction, this is subsequently marked across a 5-second window relating to the maximum duration of an initiation. This happens within coding windows (see Figure 3). Each coding window represents a 1-second block of the interaction. To obtain the 5-second interaction window, simply merge 5 windows together.

Activate interaction window (turning it blue) > right click and select ‘merge with next annotation’ > repeat until you have 5-seconds. Or using the keyboard shortcuts (much easier way to code), activate interaction window (turning it blue) and then press ‘ctrl + .’ (control + dot key). 

Figure 3: Image depicting merged initiation window of 5-seconds.

4. Coding response and verifications

After an initiation has been identified, the response and verification elements of joint attention are coded. This happens within coding windows (see Figure 3). Each coding window represents a 1-second block of the interaction. Thus, each window within the initiation periods needs to be individually evaluated and labelled for the presence or absence of a response or verification.

To do this, double click on the interaction window which will bring up a selection menu. Select the correct label for the presence or absence of response or gesture. Alternatively, click on the interaction window to highlight and use the shortcut key: ctrl + alt + m.

Figure 4: Image depicting label menu.

4. Coding episodes of shared attention

After an initiation, response and verification have been identified and coded, the joint attention episode is coded as fail or success. This is in instances whereby there is the presence of both a response and verification. Remember, if verification was coded as ‘no’, then the joint attention bout is not code.

To code the episode of shared attention following the infants response, identify the time window in which the infant engages in shared attention (see coding manual for specific criteria) and merge the windows together to reflect this

Merging windows: Activate interaction window (turning it blue) > right click and select ‘merge with next annotation’ > repeat until you have 5-seconds. Or using the keyboard shortcuts (much easier way to code), activate interaction window (turning it blue) and then press ‘ctrl + .’ (control + dot key).

Label merged window: Next double click on the merged interaction window which will bring up a selection menu. Select the correct label for the success or fail of the bid. Alternatively, click on the interaction window to highlight and use the shortcut key: ctrl + alt + m.